Parenting Project Attitudes Questionnaire Experiment 1
Emily & Mike
2015-11-03
Data analysis of basic parenting values/attitudes survey, version 1.
1 Data preprocessing
Preliminaries.
## [1] "dplyr" "langcog" "tidyr" "ggplot2" "lme4"##
## Attaching package: 'langcog'
##
## The following object is masked from 'package:base':
##
## scale
##
##
## Attaching package: 'dplyr'
##
## The following objects are masked from 'package:stats':
##
## filter, lag
##
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionRead in files and consolidate to the same directory.
files <- dir("../production-results/e1/")
d.raw <- data.frame()
for (f in files) {
jf <- paste("../production-results/e1/",f,sep="")
jd <- fromJSON(paste(readLines(jf), collapse=""))
id <- data.frame(workerid = jd$WorkerId,
sent = jd$answers$data$sent,
rating = as.numeric(jd$answers$data$rating),
children = jd$answers$data$children,
language = jd$answer$data$language)
d.raw <- bind_rows(d.raw, id)
}Note there is a repeat item here, “abstract concepts from good behavior.”
Map on question short forms so that we can use these instead.
labels <- read.csv("sent_forms.csv")
labels$sent <- as.character(labels$sent)
d.raw$sent <- as.character(d.raw$sent)
d.raw$sent <- str_replace_all(d.raw$sent, "'", "")
d.raw$sent <- str_replace_all(d.raw$sent, "’", "")
d.raw$sent <- str_replace_all(d.raw$sent, "“", "")
d.raw$sent <- str_replace_all(d.raw$sent, "”", "")
d.raw$sent <- str_replace_all(d.raw$sent, "‘", "")
d <- left_join(d.raw, labels) %>%
mutate(children = grepl("yes", children, ignore.case=TRUE))2 Basic analyses
Now look at mean ratings across sentences.
ms <- d %>%
group_by(category, instrument, short_sent) %>%
multi_boot_standard(col = "rating") %>%
arrange(instrument, category, desc(mean))
ms$short_sent_ord <- factor(ms$short_sent,
levels = ms$short_sent)
qplot(short_sent_ord, mean, col = category,
ymin = ci_lower, ymax = ci_upper,
geom = "pointrange",
data = ms) +
facet_grid(.~instrument, scale="free_x") +
theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = .5)) +
xlab("") +
ylab("Mean Rating") +
ylim(c(1,7)) +
scale_colour_solarized()
And by whether they have kids. Some interesting stuff here: the math items seem to differentiate most (along with the “newborn ignorance” question).
ms.kids <- d %>%
group_by(instrument, category, short_sent, children) %>%
multi_boot_standard(col = "rating") %>%
arrange(instrument, category, desc(mean))
ms.kids$short_sent_ord <- factor(ms.kids$short_sent,
levels = ms$short_sent)
ggplot(ms.kids, aes(x = short_sent_ord, y = mean)) +
geom_line(aes(group = children, lty = children)) +
geom_pointrange(aes(ymin = ci_lower, ymax = ci_upper,
col = category, group = short_sent_ord),
position = position_dodge(width = .3)) +
facet_grid(.~instrument, scale="free_x") +
theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = .5)) +
xlab("") +
ylab("Mean Rating") +
ylim(c(1,7)) +
scale_colour_solarized()
3 Scale reliability
3.1 Full survey
Let’s start by looking at the internal consistency of the full survey.
library(psych)
wide.all <- d %>%
select(workerid, short_sent, rating) %>%
spread(short_sent, rating)
alpha.mat <- as.matrix(select(wide.all, -workerid))
alpha(x = alpha.mat)## Some items ( baby pattern recognition baby pleasure books before reading books for babies early abstraction educational tv good examples joint attention kids quick on questions learning from senses no prediction parents reading trouble piaget pointing is good shape talk short tall stage theory talk school ) were negatively correlated with the total scale and probably should be reversed. To do this, run the function again with the 'check.keys=TRUE' option##
## Reliability analysis
## Call: alpha(x = alpha.mat)
##
## raw_alpha std.alpha G6(smc) average_r S/N ase mean sd
## 0.75 0.74 0.97 0.062 2.9 0.055 4.5 0.45
##
## lower alpha upper 95% confidence boundaries
## 0.64 0.75 0.85
##
## Reliability if an item is dropped:
## raw_alpha std.alpha G6(smc) average_r S/N
## attention important 0.74 0.74 0.97 0.061 2.8
## babies have expectations 0.74 0.74 0.97 0.063 2.9
## baby pattern recognition 0.74 0.74 0.97 0.062 2.9
## baby pleasure 0.74 0.74 0.97 0.061 2.8
## babytalk 0.74 0.74 0.97 0.062 2.8
## behavior control 0.74 0.73 0.97 0.060 2.7
## books before reading 0.75 0.74 0.97 0.062 2.9
## books for babies 0.75 0.74 0.97 0.063 2.9
## counting before math 0.72 0.73 0.97 0.058 2.7
## counting toddlers 0.74 0.74 0.97 0.061 2.8
## early abstraction 0.75 0.75 0.97 0.065 3.0
## educational tv good 0.74 0.73 0.97 0.060 2.7
## examples 0.74 0.74 0.97 0.062 2.8
## impulse control 0.74 0.74 0.97 0.061 2.8
## joint attention 0.75 0.74 0.97 0.063 2.9
## kids quick on questions 0.75 0.75 0.97 0.064 2.9
## learning from senses 0.74 0.74 0.97 0.062 2.9
## learning from surprise 0.74 0.73 0.97 0.061 2.8
## learning similar 0.74 0.74 0.97 0.061 2.8
## listen quietly 0.73 0.73 0.97 0.059 2.7
## newborn ignorance 0.75 0.75 0.97 0.064 2.9
## no baby conversation 0.73 0.74 0.97 0.061 2.8
## no babytalk 0.74 0.74 0.97 0.061 2.8
## no learning from warmth 0.74 0.74 0.97 0.063 2.9
## no prediction 0.76 0.75 0.97 0.066 3.0
## no repetition 0.75 0.75 0.97 0.066 3.0
## no talk IQ connection 0.75 0.75 0.97 0.066 3.0
## no vocab bad in school 0.73 0.73 0.97 0.059 2.7
## parents need big words 0.74 0.74 0.97 0.062 2.8
## parents reading trouble 0.75 0.75 0.97 0.066 3.0
## piaget 0.74 0.74 0.97 0.061 2.8
## pointing is good 0.74 0.73 0.97 0.060 2.7
## read all words 0.73 0.73 0.97 0.060 2.7
## reward is good 0.74 0.74 0.97 0.062 2.8
## school math 0.73 0.74 0.97 0.061 2.8
## school responsibility 0.74 0.74 0.97 0.063 2.9
## shape talk 0.74 0.74 0.97 0.062 2.8
## short tall 0.75 0.74 0.97 0.064 2.9
## stage theory 0.75 0.74 0.97 0.063 2.9
## talk school 0.75 0.74 0.97 0.063 2.9
## tell kids what to do 0.73 0.73 0.97 0.060 2.8
## tv conversation 0.74 0.74 0.97 0.061 2.8
## tv good for words 0.73 0.73 0.97 0.059 2.7
## words predict academic 0.74 0.74 0.97 0.061 2.8
## alpha se
## attention important 0.057
## babies have expectations 0.056
## baby pattern recognition 0.055
## baby pleasure 0.056
## babytalk 0.056
## behavior control 0.057
## books before reading 0.055
## books for babies 0.055
## counting before math 0.060
## counting toddlers 0.057
## early abstraction 0.054
## educational tv good 0.057
## examples 0.056
## impulse control 0.057
## joint attention 0.055
## kids quick on questions 0.055
## learning from senses 0.055
## learning from surprise 0.057
## learning similar 0.057
## listen quietly 0.058
## newborn ignorance 0.055
## no baby conversation 0.058
## no babytalk 0.057
## no learning from warmth 0.056
## no prediction 0.053
## no repetition 0.054
## no talk IQ connection 0.054
## no vocab bad in school 0.059
## parents need big words 0.057
## parents reading trouble 0.054
## piaget 0.056
## pointing is good 0.057
## read all words 0.058
## reward is good 0.056
## school math 0.058
## school responsibility 0.056
## shape talk 0.056
## short tall 0.055
## stage theory 0.055
## talk school 0.055
## tell kids what to do 0.058
## tv conversation 0.057
## tv good for words 0.058
## words predict academic 0.057
##
## Item statistics
## n raw.r std.r r.cor r.drop mean sd
## attention important 50 0.3707 0.3534 0.3446 0.291 4.8 1.7
## babies have expectations 50 0.2750 0.2426 0.2310 0.188 3.1 1.8
## baby pattern recognition 50 0.1665 0.2673 0.2615 0.115 5.8 1.0
## baby pleasure 50 0.3034 0.3555 0.3507 0.239 5.6 1.4
## babytalk 50 0.3031 0.3085 0.3081 0.216 3.8 1.8
## behavior control 50 0.4202 0.4347 0.4321 0.358 4.7 1.4
## books before reading 50 0.1902 0.2618 0.2587 0.113 5.3 1.6
## books for babies 50 0.1678 0.1912 0.1779 0.096 5.2 1.4
## counting before math 50 0.6127 0.5608 0.5595 0.540 4.5 2.1
## counting toddlers 50 0.4111 0.3459 0.3390 0.333 3.5 1.8
## early abstraction 50 -0.0055 0.0911 0.0800 -0.065 5.6 1.2
## educational tv good 50 0.3851 0.4158 0.4148 0.323 5.6 1.4
## examples 50 0.2088 0.3061 0.3049 0.153 5.9 1.1
## impulse control 50 0.3923 0.3557 0.3552 0.330 4.4 1.4
## joint attention 50 0.0988 0.1835 0.1786 0.036 5.9 1.2
## kids quick on questions 50 0.1341 0.1409 0.1370 0.055 3.3 1.6
## learning from senses 50 0.1739 0.2670 0.2648 0.120 5.8 1.1
## learning from surprise 50 0.3677 0.3819 0.3793 0.313 4.8 1.2
## learning similar 50 0.3750 0.3742 0.3667 0.310 4.6 1.4
## listen quietly 50 0.5002 0.4630 0.4601 0.442 3.9 1.5
## newborn ignorance 50 0.1958 0.1600 0.1564 0.113 4.3 1.7
## no baby conversation 50 0.4597 0.3647 0.3659 0.386 2.8 1.7
## no babytalk 50 0.3798 0.3586 0.3577 0.285 3.5 2.1
## no learning from warmth 50 0.3127 0.2112 0.2110 0.247 2.2 1.4
## no prediction 50 -0.0068 -0.0144 -0.0200 -0.107 4.8 2.0
## no repetition 50 0.0546 -0.0184 -0.0235 -0.031 2.5 1.7
## no talk IQ connection 50 0.0385 -0.0182 -0.0252 -0.041 1.9 1.6
## no vocab bad in school 50 0.5255 0.4878 0.4897 0.457 4.5 1.8
## parents need big words 50 0.3782 0.2831 0.2842 0.303 2.3 1.7
## parents reading trouble 50 -0.0596 0.0025 -0.0052 -0.130 5.3 1.4
## piaget 50 0.2662 0.3562 0.3570 0.203 5.6 1.3
## pointing is good 50 0.3658 0.4432 0.4336 0.318 5.6 1.1
## read all words 50 0.4588 0.4570 0.4550 0.389 5.4 1.7
## reward is good 50 0.2864 0.2892 0.2738 0.218 4.5 1.4
## school math 50 0.4478 0.3687 0.3696 0.377 2.7 1.7
## school responsibility 50 0.2979 0.2307 0.2284 0.205 4.3 2.0
## shape talk 50 0.2466 0.2842 0.2789 0.163 4.8 1.7
## short tall 50 0.1093 0.1704 0.1648 0.041 4.9 1.4
## stage theory 50 0.1473 0.2242 0.2222 0.083 5.7 1.3
## talk school 50 0.0963 0.1945 0.1933 0.030 6.0 1.3
## tell kids what to do 50 0.4674 0.3973 0.3854 0.393 3.1 1.8
## tv conversation 50 0.3381 0.3471 0.3429 0.279 5.4 1.3
## tv good for words 50 0.5111 0.4724 0.4754 0.450 3.6 1.5
## words predict academic 50 0.3668 0.3400 0.3407 0.279 4.2 1.9
##
## Non missing response frequency for each item
## 1 2 3 4 5 6 7 miss
## attention important 0.02 0.14 0.06 0.20 0.12 0.28 0.18 0
## babies have expectations 0.24 0.26 0.12 0.10 0.16 0.10 0.02 0
## baby pattern recognition 0.00 0.00 0.00 0.16 0.14 0.40 0.30 0
## baby pleasure 0.02 0.02 0.06 0.06 0.20 0.40 0.24 0
## babytalk 0.14 0.14 0.14 0.22 0.10 0.22 0.04 0
## behavior control 0.04 0.04 0.08 0.26 0.24 0.28 0.06 0
## books before reading 0.00 0.06 0.10 0.14 0.20 0.20 0.30 0
## books for babies 0.00 0.08 0.02 0.16 0.28 0.24 0.22 0
## counting before math 0.10 0.12 0.14 0.14 0.12 0.12 0.26 0
## counting toddlers 0.16 0.22 0.10 0.22 0.16 0.10 0.04 0
## early abstraction 0.00 0.00 0.08 0.08 0.22 0.36 0.26 0
## educational tv good 0.00 0.04 0.04 0.10 0.26 0.20 0.36 0
## examples 0.00 0.00 0.04 0.10 0.18 0.32 0.36 0
## impulse control 0.02 0.10 0.08 0.30 0.28 0.16 0.06 0
## joint attention 0.00 0.04 0.00 0.12 0.06 0.42 0.36 0
## kids quick on questions 0.16 0.16 0.26 0.16 0.16 0.10 0.00 0
## learning from senses 0.00 0.00 0.02 0.10 0.28 0.26 0.34 0
## learning from surprise 0.00 0.06 0.06 0.28 0.28 0.28 0.04 0
## learning similar 0.02 0.10 0.08 0.24 0.28 0.22 0.06 0
## listen quietly 0.08 0.04 0.24 0.36 0.12 0.12 0.04 0
## newborn ignorance 0.06 0.12 0.08 0.24 0.28 0.10 0.12 0
## no baby conversation 0.26 0.30 0.18 0.10 0.00 0.14 0.02 0
## no babytalk 0.18 0.26 0.12 0.16 0.06 0.06 0.16 0
## no learning from warmth 0.42 0.26 0.14 0.10 0.04 0.04 0.00 0
## no prediction 0.08 0.08 0.14 0.08 0.22 0.10 0.30 0
## no repetition 0.42 0.16 0.18 0.12 0.04 0.04 0.04 0
## no talk IQ connection 0.66 0.10 0.08 0.06 0.02 0.08 0.00 0
## no vocab bad in school 0.08 0.06 0.14 0.14 0.24 0.22 0.12 0
## parents need big words 0.48 0.24 0.04 0.10 0.08 0.04 0.02 0
## parents reading trouble 0.02 0.02 0.10 0.08 0.28 0.32 0.18 0
## piaget 0.00 0.04 0.02 0.12 0.20 0.32 0.30 0
## pointing is good 0.00 0.00 0.06 0.08 0.26 0.42 0.18 0
## read all words 0.02 0.04 0.10 0.12 0.14 0.20 0.38 0
## reward is good 0.04 0.08 0.06 0.28 0.34 0.12 0.08 0
## school math 0.28 0.24 0.22 0.10 0.08 0.04 0.04 0
## school responsibility 0.14 0.10 0.08 0.12 0.18 0.30 0.08 0
## shape talk 0.06 0.06 0.10 0.10 0.30 0.20 0.18 0
## short tall 0.02 0.06 0.04 0.22 0.30 0.28 0.08 0
## stage theory 0.00 0.04 0.00 0.12 0.22 0.26 0.36 0
## talk school 0.00 0.04 0.02 0.08 0.10 0.28 0.48 0
## tell kids what to do 0.20 0.32 0.10 0.14 0.10 0.12 0.02 0
## tv conversation 0.04 0.00 0.02 0.10 0.30 0.40 0.14 0
## tv good for words 0.12 0.16 0.12 0.34 0.16 0.08 0.02 0
## words predict academic 0.14 0.10 0.10 0.14 0.18 0.28 0.06 0So base \(\alpha\) = .75, which is not bad all things considered. But some items are reverse-coded, so let’s deal with that.
alpha(x = alpha.mat,
keys = labels$short_sent[labels$reverse_code==1])## Some items ( baby pattern recognition baby pleasure books before reading books for babies early abstraction educational tv good examples joint attention kids quick on questions learning from senses no prediction parents reading trouble piaget pointing is good shape talk short tall stage theory talk school ) were negatively correlated with the total scale and probably should be reversed. To do this, run the function again with the 'check.keys=TRUE' option##
## Reliability analysis
## Call: alpha(x = alpha.mat, keys = labels$short_sent[labels$reverse_code ==
## 1])
##
## raw_alpha std.alpha G6(smc) average_r S/N ase mean sd
## 0.72 0.76 0.97 0.066 3.1 0.06 4.8 0.43
##
## lower alpha upper 95% confidence boundaries
## 0.61 0.72 0.84
##
## Reliability if an item is dropped:
## raw_alpha std.alpha G6(smc) average_r S/N
## attention important 0.73 0.77 0.97 0.071 3.3
## babies have expectations 0.74 0.77 0.97 0.073 3.4
## baby pattern recognition 0.71 0.74 0.97 0.062 2.8
## baby pleasure 0.72 0.75 0.97 0.066 3.0
## babytalk 0.73 0.76 0.97 0.069 3.2
## behavior control 0.72 0.76 0.97 0.068 3.1
## books before reading 0.71 0.74 0.97 0.063 2.9
## books for babies 0.72 0.75 0.97 0.067 3.1
## counting before math- 0.72 0.76 0.97 0.068 3.1
## counting toddlers- 0.71 0.75 0.97 0.065 3.0
## early abstraction 0.71 0.74 0.97 0.063 2.9
## educational tv good- 0.74 0.77 0.97 0.073 3.4
## examples 0.71 0.74 0.97 0.062 2.8
## impulse control 0.74 0.77 0.97 0.073 3.4
## joint attention 0.70 0.74 0.97 0.062 2.8
## kids quick on questions- 0.73 0.77 0.97 0.071 3.3
## learning from senses 0.71 0.74 0.97 0.063 2.9
## learning from surprise 0.73 0.76 0.97 0.070 3.2
## learning similar 0.73 0.76 0.97 0.070 3.2
## listen quietly- 0.72 0.76 0.97 0.067 3.1
## newborn ignorance- 0.72 0.76 0.97 0.067 3.1
## no baby conversation- 0.70 0.74 0.97 0.062 2.8
## no babytalk- 0.72 0.76 0.97 0.069 3.2
## no learning from warmth- 0.70 0.74 0.97 0.061 2.8
## no prediction- 0.73 0.76 0.97 0.069 3.2
## no repetition- 0.71 0.75 0.97 0.064 2.9
## no talk IQ connection- 0.71 0.75 0.97 0.065 3.0
## no vocab bad in school 0.74 0.77 0.97 0.071 3.3
## parents need big words- 0.70 0.74 0.97 0.062 2.8
## parents reading trouble 0.72 0.75 0.97 0.066 3.0
## piaget 0.70 0.74 0.97 0.062 2.8
## pointing is good 0.71 0.74 0.97 0.063 2.9
## read all words- 0.74 0.77 0.97 0.073 3.4
## reward is good 0.73 0.77 0.97 0.070 3.3
## school math- 0.70 0.74 0.97 0.063 2.9
## school responsibility- 0.72 0.75 0.97 0.066 3.0
## shape talk 0.71 0.75 0.97 0.065 3.0
## short tall 0.71 0.75 0.97 0.065 3.0
## stage theory 0.71 0.74 0.97 0.063 2.9
## talk school 0.70 0.74 0.97 0.062 2.8
## tell kids what to do- 0.71 0.75 0.97 0.064 3.0
## tv conversation 0.73 0.76 0.97 0.070 3.2
## tv good for words- 0.71 0.75 0.97 0.066 3.1
## words predict academic 0.73 0.76 0.97 0.069 3.2
## alpha se
## attention important 0.057
## babies have expectations 0.056
## baby pattern recognition 0.063
## baby pleasure 0.061
## babytalk 0.059
## behavior control 0.060
## books before reading 0.063
## books for babies 0.061
## counting before math- 0.060
## counting toddlers- 0.063
## early abstraction 0.063
## educational tv good- 0.057
## examples 0.063
## impulse control 0.057
## joint attention 0.064
## kids quick on questions- 0.058
## learning from senses 0.062
## learning from surprise 0.059
## learning similar 0.058
## listen quietly- 0.061
## newborn ignorance- 0.061
## no baby conversation- 0.065
## no babytalk- 0.059
## no learning from warmth- 0.065
## no prediction- 0.059
## no repetition- 0.063
## no talk IQ connection- 0.062
## no vocab bad in school 0.057
## parents need big words- 0.064
## parents reading trouble 0.061
## piaget 0.063
## pointing is good 0.062
## read all words- 0.056
## reward is good 0.058
## school math- 0.064
## school responsibility- 0.061
## shape talk 0.062
## short tall 0.061
## stage theory 0.063
## talk school 0.063
## tell kids what to do- 0.063
## tv conversation 0.059
## tv good for words- 0.061
## words predict academic 0.058
##
## Item statistics
## n raw.r std.r r.cor r.drop mean sd
## attention important 50 -0.0182 0.0093 -0.0040 -0.109 4.8 1.7
## babies have expectations 50 -0.1762 -0.1718 -0.1886 -0.265 3.1 1.8
## baby pattern recognition 50 0.6026 0.6235 0.6226 0.566 5.8 1.0
## baby pleasure 50 0.3078 0.3447 0.3397 0.240 5.6 1.4
## babytalk 50 0.1251 0.1304 0.1275 0.030 3.8 1.8
## behavior control 50 0.1559 0.1936 0.1877 0.082 4.7 1.4
## books before reading 50 0.5112 0.5275 0.5280 0.447 5.3 1.6
## books for babies 50 0.2801 0.2755 0.2635 0.209 5.2 1.4
## counting before math- 50 0.2326 0.1824 0.1759 0.126 3.5 2.1
## counting toddlers- 50 0.4683 0.4264 0.4207 0.391 4.5 1.8
## early abstraction 50 0.5113 0.5328 0.5278 0.463 5.6 1.2
## educational tv good- 50 -0.1272 -0.1724 -0.1814 -0.197 2.4 1.4
## examples 50 0.5715 0.6082 0.6109 0.529 5.9 1.1
## impulse control 50 -0.2089 -0.1885 -0.1964 -0.277 4.4 1.4
## joint attention 50 0.6091 0.6278 0.6289 0.565 5.9 1.2
## kids quick on questions- 50 -0.0345 -0.0284 -0.0345 -0.116 4.7 1.6
## learning from senses 50 0.5090 0.5427 0.5441 0.465 5.8 1.1
## learning from surprise 50 0.0244 0.0637 0.0568 -0.039 4.8 1.2
## learning similar 50 -0.0042 0.0280 0.0159 -0.080 4.6 1.4
## listen quietly- 50 0.2888 0.2545 0.2487 0.217 4.1 1.5
## newborn ignorance- 50 0.2681 0.2528 0.2505 0.184 3.7 1.7
## no baby conversation- 50 0.6294 0.5978 0.6020 0.569 5.2 1.7
## no babytalk- 50 0.2144 0.1472 0.1433 0.107 4.5 2.1
## no learning from warmth- 50 0.6780 0.6730 0.6788 0.635 5.8 1.4
## no prediction- 50 0.1700 0.1346 0.1312 0.066 3.2 2.0
## no repetition- 50 0.4603 0.4787 0.4803 0.385 5.5 1.7
## no talk IQ connection- 50 0.3821 0.4063 0.4052 0.308 6.1 1.6
## no vocab bad in school 50 -0.0513 -0.0422 -0.0476 -0.142 4.5 1.8
## parents need big words- 50 0.6020 0.5933 0.5984 0.542 5.7 1.7
## parents reading trouble 50 0.3126 0.3329 0.3297 0.244 5.3 1.4
## piaget 50 0.5567 0.5989 0.6028 0.507 5.6 1.3
## pointing is good 50 0.5129 0.5527 0.5447 0.470 5.6 1.1
## read all words- 50 -0.1436 -0.1712 -0.1817 -0.227 2.6 1.7
## reward is good 50 0.0083 0.0175 -0.0012 -0.067 4.5 1.4
## school math- 50 0.5961 0.5539 0.5572 0.535 5.3 1.7
## school responsibility- 50 0.3270 0.3136 0.3124 0.231 3.7 2.0
## shape talk 50 0.3750 0.3640 0.3599 0.294 4.8 1.7
## short tall 50 0.3604 0.3732 0.3704 0.296 4.9 1.4
## stage theory 50 0.5138 0.5387 0.5408 0.461 5.7 1.3
## talk school 50 0.5805 0.6126 0.6169 0.532 6.0 1.3
## tell kids what to do- 50 0.4883 0.4388 0.4276 0.413 4.9 1.8
## tv conversation 50 0.0263 0.0602 0.0522 -0.041 5.4 1.3
## tv good for words- 50 0.3306 0.2931 0.2935 0.256 4.4 1.5
## words predict academic 50 0.1158 0.0940 0.0913 0.016 4.2 1.9
##
## Non missing response frequency for each item
## 1 2 3 4 5 6 7 miss
## attention important 0.02 0.14 0.06 0.20 0.12 0.28 0.18 0
## babies have expectations 0.24 0.26 0.12 0.10 0.16 0.10 0.02 0
## baby pattern recognition 0.00 0.00 0.00 0.16 0.14 0.40 0.30 0
## baby pleasure 0.02 0.02 0.06 0.06 0.20 0.40 0.24 0
## babytalk 0.14 0.14 0.14 0.22 0.10 0.22 0.04 0
## behavior control 0.04 0.04 0.08 0.26 0.24 0.28 0.06 0
## books before reading 0.00 0.06 0.10 0.14 0.20 0.20 0.30 0
## books for babies 0.00 0.08 0.02 0.16 0.28 0.24 0.22 0
## counting before math 0.10 0.12 0.14 0.14 0.12 0.12 0.26 0
## counting toddlers 0.16 0.22 0.10 0.22 0.16 0.10 0.04 0
## early abstraction 0.00 0.00 0.08 0.08 0.22 0.36 0.26 0
## educational tv good 0.00 0.04 0.04 0.10 0.26 0.20 0.36 0
## examples 0.00 0.00 0.04 0.10 0.18 0.32 0.36 0
## impulse control 0.02 0.10 0.08 0.30 0.28 0.16 0.06 0
## joint attention 0.00 0.04 0.00 0.12 0.06 0.42 0.36 0
## kids quick on questions 0.16 0.16 0.26 0.16 0.16 0.10 0.00 0
## learning from senses 0.00 0.00 0.02 0.10 0.28 0.26 0.34 0
## learning from surprise 0.00 0.06 0.06 0.28 0.28 0.28 0.04 0
## learning similar 0.02 0.10 0.08 0.24 0.28 0.22 0.06 0
## listen quietly 0.08 0.04 0.24 0.36 0.12 0.12 0.04 0
## newborn ignorance 0.06 0.12 0.08 0.24 0.28 0.10 0.12 0
## no baby conversation 0.26 0.30 0.18 0.10 0.00 0.14 0.02 0
## no babytalk 0.18 0.26 0.12 0.16 0.06 0.06 0.16 0
## no learning from warmth 0.42 0.26 0.14 0.10 0.04 0.04 0.00 0
## no prediction 0.08 0.08 0.14 0.08 0.22 0.10 0.30 0
## no repetition 0.42 0.16 0.18 0.12 0.04 0.04 0.04 0
## no talk IQ connection 0.66 0.10 0.08 0.06 0.02 0.08 0.00 0
## no vocab bad in school 0.08 0.06 0.14 0.14 0.24 0.22 0.12 0
## parents need big words 0.48 0.24 0.04 0.10 0.08 0.04 0.02 0
## parents reading trouble 0.02 0.02 0.10 0.08 0.28 0.32 0.18 0
## piaget 0.00 0.04 0.02 0.12 0.20 0.32 0.30 0
## pointing is good 0.00 0.00 0.06 0.08 0.26 0.42 0.18 0
## read all words 0.02 0.04 0.10 0.12 0.14 0.20 0.38 0
## reward is good 0.04 0.08 0.06 0.28 0.34 0.12 0.08 0
## school math 0.28 0.24 0.22 0.10 0.08 0.04 0.04 0
## school responsibility 0.14 0.10 0.08 0.12 0.18 0.30 0.08 0
## shape talk 0.06 0.06 0.10 0.10 0.30 0.20 0.18 0
## short tall 0.02 0.06 0.04 0.22 0.30 0.28 0.08 0
## stage theory 0.00 0.04 0.00 0.12 0.22 0.26 0.36 0
## talk school 0.00 0.04 0.02 0.08 0.10 0.28 0.48 0
## tell kids what to do 0.20 0.32 0.10 0.14 0.10 0.12 0.02 0
## tv conversation 0.04 0.00 0.02 0.10 0.30 0.40 0.14 0
## tv good for words 0.12 0.16 0.12 0.34 0.16 0.08 0.02 0
## words predict academic 0.14 0.10 0.10 0.14 0.18 0.28 0.06 0Actually, this is worse than if we do things automatically, which yields a higher \(\alpha = .85\), though this could well be overfitting.
alpha(x = alpha.mat,
check.keys = TRUE)##
## Reliability analysis
## Call: alpha(x = alpha.mat, check.keys = TRUE)
##
## raw_alpha std.alpha G6(smc) average_r S/N ase mean sd
## 0.85 0.85 0.98 0.12 5.9 0.035 3.3 0.57
##
## lower alpha upper 95% confidence boundaries
## 0.78 0.85 0.92
##
## Reliability if an item is dropped:
## raw_alpha std.alpha G6(smc) average_r S/N
## attention important 0.85 0.86 0.98 0.12 5.9
## babies have expectations 0.85 0.85 0.98 0.12 5.8
## baby pattern recognition- 0.84 0.85 0.98 0.12 5.6
## baby pleasure- 0.85 0.85 0.98 0.12 5.9
## babytalk 0.85 0.86 0.98 0.12 6.0
## behavior control 0.85 0.86 0.98 0.12 6.0
## books before reading- 0.84 0.85 0.98 0.12 5.6
## books for babies- 0.85 0.86 0.98 0.12 5.9
## counting before math 0.84 0.85 0.98 0.11 5.6
## counting toddlers 0.84 0.85 0.98 0.11 5.6
## early abstraction- 0.84 0.85 0.98 0.11 5.5
## educational tv good- 0.85 0.86 0.98 0.12 6.0
## examples- 0.85 0.85 0.98 0.12 5.7
## impulse control 0.84 0.85 0.98 0.12 5.6
## joint attention- 0.84 0.85 0.98 0.12 5.7
## kids quick on questions- 0.85 0.86 0.98 0.12 6.0
## learning from senses- 0.84 0.85 0.98 0.12 5.6
## learning from surprise 0.85 0.85 0.98 0.12 5.9
## learning similar 0.85 0.86 0.98 0.12 5.9
## listen quietly 0.84 0.85 0.98 0.12 5.7
## newborn ignorance 0.85 0.85 0.98 0.12 5.7
## no baby conversation 0.84 0.84 0.98 0.11 5.4
## no babytalk 0.85 0.86 0.98 0.12 6.0
## no learning from warmth 0.84 0.84 0.98 0.11 5.4
## no prediction- 0.85 0.86 0.98 0.12 5.9
## no repetition 0.84 0.85 0.98 0.12 5.7
## no talk IQ connection 0.85 0.85 0.98 0.12 5.8
## no vocab bad in school 0.84 0.85 0.98 0.12 5.6
## parents need big words 0.84 0.85 0.98 0.11 5.5
## parents reading trouble- 0.84 0.85 0.98 0.12 5.6
## piaget- 0.85 0.85 0.98 0.12 5.7
## pointing is good- 0.85 0.85 0.98 0.12 5.8
## read all words 0.85 0.85 0.98 0.12 5.9
## reward is good 0.85 0.86 0.98 0.12 5.9
## school math 0.84 0.84 0.98 0.11 5.4
## school responsibility 0.84 0.85 0.98 0.12 5.7
## shape talk- 0.85 0.86 0.98 0.12 5.9
## short tall- 0.85 0.85 0.98 0.12 5.7
## stage theory- 0.84 0.85 0.98 0.12 5.7
## talk school- 0.84 0.85 0.98 0.11 5.5
## tell kids what to do 0.84 0.85 0.98 0.11 5.6
## tv conversation 0.85 0.86 0.98 0.12 5.9
## tv good for words 0.84 0.85 0.98 0.12 5.6
## words predict academic 0.85 0.85 0.98 0.12 5.9
## alpha se
## attention important 0.035
## babies have expectations 0.035
## baby pattern recognition- 0.036
## baby pleasure- 0.035
## babytalk 0.034
## behavior control 0.035
## books before reading- 0.036
## books for babies- 0.035
## counting before math 0.037
## counting toddlers 0.037
## early abstraction- 0.036
## educational tv good- 0.034
## examples- 0.036
## impulse control 0.036
## joint attention- 0.036
## kids quick on questions- 0.034
## learning from senses- 0.036
## learning from surprise 0.035
## learning similar 0.035
## listen quietly 0.036
## newborn ignorance 0.036
## no baby conversation 0.037
## no babytalk 0.034
## no learning from warmth 0.037
## no prediction- 0.035
## no repetition 0.036
## no talk IQ connection 0.035
## no vocab bad in school 0.036
## parents need big words 0.037
## parents reading trouble- 0.036
## piaget- 0.035
## pointing is good- 0.035
## read all words 0.035
## reward is good 0.035
## school math 0.037
## school responsibility 0.036
## shape talk- 0.035
## short tall- 0.036
## stage theory- 0.036
## talk school- 0.036
## tell kids what to do 0.037
## tv conversation 0.035
## tv good for words 0.036
## words predict academic 0.035
##
## Item statistics
## n raw.r std.r r.cor r.drop mean sd
## attention important 50 0.219 0.191 0.182 0.151 4.8 1.7
## babies have expectations 50 0.292 0.274 0.264 0.224 3.1 1.8
## baby pattern recognition- 50 0.447 0.495 0.492 0.413 2.2 1.0
## baby pleasure- 50 0.192 0.239 0.233 0.138 2.4 1.4
## babytalk 50 0.156 0.115 0.112 0.084 3.8 1.8
## behavior control 50 0.183 0.157 0.152 0.127 4.7 1.4
## books before reading- 50 0.491 0.523 0.522 0.441 2.7 1.6
## books for babies- 50 0.166 0.184 0.173 0.109 2.8 1.4
## counting before math 50 0.583 0.538 0.535 0.523 4.5 2.1
## counting toddlers 50 0.556 0.547 0.542 0.504 3.5 1.8
## early abstraction- 50 0.525 0.562 0.557 0.489 2.4 1.2
## educational tv good- 50 0.041 0.069 0.064 -0.014 2.4 1.4
## examples- 50 0.375 0.415 0.414 0.334 2.1 1.1
## impulse control 50 0.533 0.524 0.524 0.491 4.4 1.4
## joint attention- 50 0.399 0.441 0.439 0.356 2.1 1.2
## kids quick on questions- 50 0.097 0.101 0.097 0.035 4.7 1.6
## learning from senses- 50 0.426 0.471 0.470 0.389 2.2 1.1
## learning from surprise 50 0.237 0.218 0.214 0.191 4.8 1.2
## learning similar 50 0.189 0.171 0.163 0.133 4.6 1.4
## listen quietly 50 0.418 0.395 0.391 0.368 3.9 1.5
## newborn ignorance 50 0.384 0.386 0.384 0.325 4.3 1.7
## no baby conversation 50 0.673 0.667 0.669 0.633 2.8 1.7
## no babytalk 50 0.165 0.130 0.126 0.083 3.5 2.1
## no learning from warmth 50 0.714 0.720 0.723 0.685 2.2 1.4
## no prediction- 50 0.207 0.187 0.184 0.129 3.2 2.0
## no repetition 50 0.420 0.435 0.435 0.361 2.5 1.7
## no talk IQ connection 50 0.279 0.307 0.304 0.219 1.9 1.6
## no vocab bad in school 50 0.510 0.467 0.467 0.456 4.5 1.8
## parents need big words 50 0.644 0.644 0.647 0.602 2.3 1.7
## parents reading trouble- 50 0.487 0.491 0.489 0.442 2.7 1.4
## piaget- 50 0.312 0.368 0.368 0.264 2.4 1.3
## pointing is good- 50 0.239 0.294 0.284 0.198 2.4 1.1
## read all words 50 0.267 0.216 0.211 0.203 5.4 1.7
## reward is good 50 0.207 0.195 0.182 0.151 4.5 1.4
## school math 50 0.683 0.673 0.676 0.645 2.7 1.7
## school responsibility 50 0.478 0.457 0.456 0.414 4.3 2.0
## shape talk- 50 0.175 0.200 0.195 0.107 3.2 1.7
## short tall- 50 0.367 0.386 0.383 0.319 3.1 1.4
## stage theory- 50 0.407 0.446 0.446 0.362 2.3 1.3
## talk school- 50 0.551 0.583 0.585 0.512 2.0 1.3
## tell kids what to do 50 0.570 0.548 0.539 0.520 3.1 1.8
## tv conversation 50 0.207 0.181 0.175 0.157 5.4 1.3
## tv good for words 50 0.527 0.499 0.500 0.480 3.6 1.5
## words predict academic 50 0.258 0.228 0.226 0.185 4.2 1.9
##
## Non missing response frequency for each item
## 1 2 3 4 5 6 7 miss
## attention important 0.02 0.14 0.06 0.20 0.12 0.28 0.18 0
## babies have expectations 0.24 0.26 0.12 0.10 0.16 0.10 0.02 0
## baby pattern recognition 0.00 0.00 0.00 0.16 0.14 0.40 0.30 0
## baby pleasure 0.02 0.02 0.06 0.06 0.20 0.40 0.24 0
## babytalk 0.14 0.14 0.14 0.22 0.10 0.22 0.04 0
## behavior control 0.04 0.04 0.08 0.26 0.24 0.28 0.06 0
## books before reading 0.00 0.06 0.10 0.14 0.20 0.20 0.30 0
## books for babies 0.00 0.08 0.02 0.16 0.28 0.24 0.22 0
## counting before math 0.10 0.12 0.14 0.14 0.12 0.12 0.26 0
## counting toddlers 0.16 0.22 0.10 0.22 0.16 0.10 0.04 0
## early abstraction 0.00 0.00 0.08 0.08 0.22 0.36 0.26 0
## educational tv good 0.00 0.04 0.04 0.10 0.26 0.20 0.36 0
## examples 0.00 0.00 0.04 0.10 0.18 0.32 0.36 0
## impulse control 0.02 0.10 0.08 0.30 0.28 0.16 0.06 0
## joint attention 0.00 0.04 0.00 0.12 0.06 0.42 0.36 0
## kids quick on questions 0.16 0.16 0.26 0.16 0.16 0.10 0.00 0
## learning from senses 0.00 0.00 0.02 0.10 0.28 0.26 0.34 0
## learning from surprise 0.00 0.06 0.06 0.28 0.28 0.28 0.04 0
## learning similar 0.02 0.10 0.08 0.24 0.28 0.22 0.06 0
## listen quietly 0.08 0.04 0.24 0.36 0.12 0.12 0.04 0
## newborn ignorance 0.06 0.12 0.08 0.24 0.28 0.10 0.12 0
## no baby conversation 0.26 0.30 0.18 0.10 0.00 0.14 0.02 0
## no babytalk 0.18 0.26 0.12 0.16 0.06 0.06 0.16 0
## no learning from warmth 0.42 0.26 0.14 0.10 0.04 0.04 0.00 0
## no prediction 0.08 0.08 0.14 0.08 0.22 0.10 0.30 0
## no repetition 0.42 0.16 0.18 0.12 0.04 0.04 0.04 0
## no talk IQ connection 0.66 0.10 0.08 0.06 0.02 0.08 0.00 0
## no vocab bad in school 0.08 0.06 0.14 0.14 0.24 0.22 0.12 0
## parents need big words 0.48 0.24 0.04 0.10 0.08 0.04 0.02 0
## parents reading trouble 0.02 0.02 0.10 0.08 0.28 0.32 0.18 0
## piaget 0.00 0.04 0.02 0.12 0.20 0.32 0.30 0
## pointing is good 0.00 0.00 0.06 0.08 0.26 0.42 0.18 0
## read all words 0.02 0.04 0.10 0.12 0.14 0.20 0.38 0
## reward is good 0.04 0.08 0.06 0.28 0.34 0.12 0.08 0
## school math 0.28 0.24 0.22 0.10 0.08 0.04 0.04 0
## school responsibility 0.14 0.10 0.08 0.12 0.18 0.30 0.08 0
## shape talk 0.06 0.06 0.10 0.10 0.30 0.20 0.18 0
## short tall 0.02 0.06 0.04 0.22 0.30 0.28 0.08 0
## stage theory 0.00 0.04 0.00 0.12 0.22 0.26 0.36 0
## talk school 0.00 0.04 0.02 0.08 0.10 0.28 0.48 0
## tell kids what to do 0.20 0.32 0.10 0.14 0.10 0.12 0.02 0
## tv conversation 0.04 0.00 0.02 0.10 0.30 0.40 0.14 0
## tv good for words 0.12 0.16 0.12 0.34 0.16 0.08 0.02 0
## words predict academic 0.14 0.10 0.10 0.14 0.18 0.28 0.06 0So now let’s do this for the sub-surveys, where these issues are maybe bit clearer.
3.2 Attitudes
First the attitudes items.
library(psych)
wide.attitudes <- d %>%
filter(instrument == "attitudes") %>%
select(workerid, short_sent, rating) %>%
spread(short_sent, rating)
alpha.mat <- as.matrix(select(wide.attitudes, -workerid))
alpha(x = alpha.mat)## Some items ( babies have expectations impulse control newborn ignorance reward is good ) were negatively correlated with the total scale and probably should be reversed. To do this, run the function again with the 'check.keys=TRUE' option##
## Reliability analysis
## Call: alpha(x = alpha.mat)
##
## raw_alpha std.alpha G6(smc) average_r S/N ase mean sd
## 0.55 0.6 0.77 0.097 1.5 0.1 5 0.52
##
## lower alpha upper 95% confidence boundaries
## 0.34 0.55 0.75
##
## Reliability if an item is dropped:
## raw_alpha std.alpha G6(smc) average_r S/N
## babies have expectations 0.60 0.63 0.78 0.116 1.7
## baby pattern recognition 0.51 0.56 0.74 0.090 1.3
## baby pleasure 0.50 0.56 0.73 0.089 1.3
## behavior control 0.49 0.56 0.72 0.089 1.3
## early abstraction 0.53 0.58 0.75 0.097 1.4
## examples 0.47 0.52 0.70 0.078 1.1
## impulse control 0.53 0.60 0.76 0.103 1.5
## learning from senses 0.52 0.57 0.73 0.092 1.3
## learning from surprise 0.50 0.57 0.75 0.092 1.3
## learning similar 0.56 0.61 0.77 0.108 1.6
## newborn ignorance 0.60 0.64 0.78 0.119 1.8
## piaget 0.47 0.53 0.72 0.081 1.1
## reward is good 0.55 0.61 0.78 0.107 1.6
## stage theory 0.53 0.58 0.75 0.097 1.4
## alpha se
## babies have expectations 0.095
## baby pattern recognition 0.111
## baby pleasure 0.114
## behavior control 0.115
## early abstraction 0.108
## examples 0.119
## impulse control 0.109
## learning from senses 0.110
## learning from surprise 0.114
## learning similar 0.103
## newborn ignorance 0.095
## piaget 0.118
## reward is good 0.104
## stage theory 0.108
##
## Item statistics
## n raw.r std.r r.cor r.drop mean sd
## babies have expectations 50 0.20 0.138 0.02606 -0.051 3.1 1.8
## baby pattern recognition 50 0.43 0.497 0.46889 0.299 5.8 1.0
## baby pleasure 50 0.50 0.517 0.49958 0.339 5.6 1.4
## behavior control 50 0.53 0.507 0.50797 0.369 4.7 1.4
## early abstraction 50 0.33 0.403 0.36556 0.178 5.6 1.2
## examples 50 0.63 0.664 0.68698 0.516 5.9 1.1
## impulse control 50 0.39 0.321 0.25022 0.206 4.4 1.4
## learning from senses 50 0.40 0.473 0.45866 0.268 5.8 1.1
## learning from surprise 50 0.50 0.474 0.41017 0.354 4.8 1.2
## learning similar 50 0.26 0.252 0.15764 0.063 4.6 1.4
## newborn ignorance 50 0.15 0.091 -0.00051 -0.084 4.3 1.7
## piaget 50 0.60 0.627 0.62091 0.463 5.6 1.3
## reward is good 50 0.30 0.257 0.14328 0.107 4.5 1.4
## stage theory 50 0.37 0.404 0.34980 0.202 5.7 1.3
##
## Non missing response frequency for each item
## 1 2 3 4 5 6 7 miss
## babies have expectations 0.24 0.26 0.12 0.10 0.16 0.10 0.02 0
## baby pattern recognition 0.00 0.00 0.00 0.16 0.14 0.40 0.30 0
## baby pleasure 0.02 0.02 0.06 0.06 0.20 0.40 0.24 0
## behavior control 0.04 0.04 0.08 0.26 0.24 0.28 0.06 0
## early abstraction 0.00 0.00 0.08 0.08 0.22 0.36 0.26 0
## examples 0.00 0.00 0.04 0.10 0.18 0.32 0.36 0
## impulse control 0.02 0.10 0.08 0.30 0.28 0.16 0.06 0
## learning from senses 0.00 0.00 0.02 0.10 0.28 0.26 0.34 0
## learning from surprise 0.00 0.06 0.06 0.28 0.28 0.28 0.04 0
## learning similar 0.02 0.10 0.08 0.24 0.28 0.22 0.06 0
## newborn ignorance 0.06 0.12 0.08 0.24 0.28 0.10 0.12 0
## piaget 0.00 0.04 0.02 0.12 0.20 0.32 0.30 0
## reward is good 0.04 0.08 0.06 0.28 0.34 0.12 0.08 0
## stage theory 0.00 0.04 0.00 0.12 0.22 0.26 0.36 0OK, so these don’t hang together very well, which is in some sense probably a good thing from the perspective of seeing multi-dimensionality in parenting.
3.3 Knowledge
Now the knowledge items. These are in fact reverse-coded.
library(psych)
wide.knowledge <- d %>%
filter(instrument == "knowledge") %>%
select(workerid, short_sent, rating) %>%
spread(short_sent, rating)
alpha.mat <- as.matrix(select(wide.knowledge, -workerid))
alpha(x = alpha.mat,
keys = labels$short_sent[labels$reverse_code == 1 &
labels$instrument == "knowledge"])## Some items ( books before reading books for babies educational tv good joint attention kids quick on questions no prediction parents reading trouble pointing is good shape talk short tall talk school ) were negatively correlated with the total scale and probably should be reversed. To do this, run the function again with the 'check.keys=TRUE' option##
## Reliability analysis
## Call: alpha(x = alpha.mat, keys = labels$short_sent[labels$reverse_code ==
## 1 & labels$instrument == "knowledge"])
##
## raw_alpha std.alpha G6(smc) average_r S/N ase mean sd
## 0.71 0.72 0.92 0.08 2.6 0.065 4.7 0.53
##
## lower alpha upper 95% confidence boundaries
## 0.58 0.71 0.83
##
## Reliability if an item is dropped:
## raw_alpha std.alpha G6(smc) average_r S/N
## attention important 0.72 0.74 0.92 0.087 2.8
## babytalk 0.72 0.74 0.92 0.088 2.8
## books before reading 0.69 0.70 0.91 0.076 2.4
## books for babies 0.70 0.72 0.92 0.081 2.6
## counting before math- 0.69 0.71 0.92 0.079 2.5
## counting toddlers- 0.68 0.70 0.92 0.075 2.3
## educational tv good- 0.72 0.74 0.92 0.088 2.8
## joint attention 0.68 0.70 0.91 0.074 2.3
## kids quick on questions- 0.71 0.73 0.92 0.086 2.7
## listen quietly- 0.69 0.71 0.91 0.078 2.5
## no baby conversation- 0.67 0.69 0.91 0.072 2.2
## no babytalk- 0.70 0.72 0.92 0.080 2.5
## no learning from warmth- 0.67 0.68 0.91 0.070 2.2
## no prediction- 0.72 0.73 0.91 0.086 2.7
## no repetition- 0.69 0.71 0.91 0.077 2.4
## no talk IQ connection- 0.70 0.71 0.92 0.079 2.5
## no vocab bad in school 0.73 0.74 0.92 0.090 2.9
## parents need big words- 0.67 0.69 0.90 0.072 2.2
## parents reading trouble 0.70 0.72 0.91 0.080 2.5
## pointing is good 0.70 0.71 0.92 0.078 2.5
## read all words- 0.73 0.74 0.93 0.091 2.9
## school math- 0.67 0.69 0.91 0.072 2.2
## school responsibility- 0.70 0.71 0.91 0.079 2.5
## shape talk 0.70 0.72 0.92 0.081 2.5
## short tall 0.70 0.71 0.92 0.079 2.5
## talk school 0.68 0.70 0.91 0.073 2.3
## tell kids what to do- 0.67 0.70 0.92 0.073 2.3
## tv conversation 0.72 0.74 0.92 0.088 2.8
## tv good for words- 0.69 0.71 0.90 0.076 2.4
## words predict academic 0.72 0.73 0.91 0.086 2.7
## alpha se
## attention important 0.062
## babytalk 0.062
## books before reading 0.069
## books for babies 0.066
## counting before math- 0.068
## counting toddlers- 0.070
## educational tv good- 0.063
## joint attention 0.069
## kids quick on questions- 0.064
## listen quietly- 0.068
## no baby conversation- 0.072
## no babytalk- 0.067
## no learning from warmth- 0.072
## no prediction- 0.063
## no repetition- 0.068
## no talk IQ connection- 0.067
## no vocab bad in school 0.061
## parents need big words- 0.072
## parents reading trouble 0.067
## pointing is good 0.067
## read all words- 0.061
## school math- 0.072
## school responsibility- 0.067
## shape talk 0.066
## short tall 0.067
## talk school 0.070
## tell kids what to do- 0.071
## tv conversation 0.063
## tv good for words- 0.069
## words predict academic 0.063
##
## Item statistics
## n raw.r std.r r.cor r.drop mean sd
## attention important 50 -0.0004 0.0226 -0.012 -0.109 4.8 1.7
## babytalk 50 0.0186 0.0077 -0.018 -0.096 3.8 1.8
## books before reading 50 0.4603 0.4851 0.478 0.378 5.3 1.6
## books for babies 50 0.2555 0.2752 0.241 0.169 5.2 1.4
## counting before math- 50 0.3933 0.3579 0.346 0.275 3.5 2.1
## counting toddlers- 50 0.5580 0.5378 0.528 0.475 4.5 1.8
## educational tv good- 50 0.0214 -0.0135 -0.038 -0.065 2.4 1.4
## joint attention 50 0.5500 0.5810 0.579 0.492 5.9 1.2
## kids quick on questions- 50 0.0719 0.0860 0.056 -0.027 4.7 1.6
## listen quietly- 50 0.3998 0.3871 0.372 0.319 4.1 1.5
## no baby conversation- 50 0.6740 0.6704 0.681 0.607 5.2 1.7
## no babytalk- 50 0.3748 0.3296 0.302 0.255 4.5 2.1
## no learning from warmth- 50 0.7258 0.7439 0.759 0.680 5.8 1.4
## no prediction- 50 0.1379 0.0992 0.089 0.013 3.2 2.0
## no repetition- 50 0.4194 0.4465 0.446 0.325 5.5 1.7
## no talk IQ connection- 50 0.3447 0.3766 0.365 0.253 6.1 1.6
## no vocab bad in school 50 -0.0909 -0.0916 -0.115 -0.198 4.5 1.8
## parents need big words- 50 0.6420 0.6526 0.666 0.574 5.7 1.7
## parents reading trouble 50 0.2982 0.3231 0.306 0.215 5.3 1.4
## pointing is good 50 0.3560 0.4002 0.378 0.295 5.6 1.1
## read all words- 50 -0.0801 -0.1113 -0.142 -0.182 2.6 1.7
## school math- 50 0.6717 0.6589 0.666 0.608 5.3 1.7
## school responsibility- 50 0.3626 0.3575 0.350 0.249 3.7 2.0
## shape talk 50 0.3044 0.2929 0.278 0.202 4.8 1.7
## short tall 50 0.3383 0.3488 0.336 0.259 4.9 1.4
## talk school 50 0.5538 0.5971 0.599 0.492 6.0 1.3
## tell kids what to do- 50 0.6190 0.5949 0.579 0.543 4.9 1.8
## tv conversation 50 -0.0227 0.0110 -0.021 -0.103 5.4 1.3
## tv good for words- 50 0.4842 0.4715 0.476 0.405 4.4 1.5
## words predict academic 50 0.0891 0.0729 0.062 -0.031 4.2 1.9
##
## Non missing response frequency for each item
## 1 2 3 4 5 6 7 miss
## attention important 0.02 0.14 0.06 0.20 0.12 0.28 0.18 0
## babytalk 0.14 0.14 0.14 0.22 0.10 0.22 0.04 0
## books before reading 0.00 0.06 0.10 0.14 0.20 0.20 0.30 0
## books for babies 0.00 0.08 0.02 0.16 0.28 0.24 0.22 0
## counting before math 0.10 0.12 0.14 0.14 0.12 0.12 0.26 0
## counting toddlers 0.16 0.22 0.10 0.22 0.16 0.10 0.04 0
## educational tv good 0.00 0.04 0.04 0.10 0.26 0.20 0.36 0
## joint attention 0.00 0.04 0.00 0.12 0.06 0.42 0.36 0
## kids quick on questions 0.16 0.16 0.26 0.16 0.16 0.10 0.00 0
## listen quietly 0.08 0.04 0.24 0.36 0.12 0.12 0.04 0
## no baby conversation 0.26 0.30 0.18 0.10 0.00 0.14 0.02 0
## no babytalk 0.18 0.26 0.12 0.16 0.06 0.06 0.16 0
## no learning from warmth 0.42 0.26 0.14 0.10 0.04 0.04 0.00 0
## no prediction 0.08 0.08 0.14 0.08 0.22 0.10 0.30 0
## no repetition 0.42 0.16 0.18 0.12 0.04 0.04 0.04 0
## no talk IQ connection 0.66 0.10 0.08 0.06 0.02 0.08 0.00 0
## no vocab bad in school 0.08 0.06 0.14 0.14 0.24 0.22 0.12 0
## parents need big words 0.48 0.24 0.04 0.10 0.08 0.04 0.02 0
## parents reading trouble 0.02 0.02 0.10 0.08 0.28 0.32 0.18 0
## pointing is good 0.00 0.00 0.06 0.08 0.26 0.42 0.18 0
## read all words 0.02 0.04 0.10 0.12 0.14 0.20 0.38 0
## school math 0.28 0.24 0.22 0.10 0.08 0.04 0.04 0
## school responsibility 0.14 0.10 0.08 0.12 0.18 0.30 0.08 0
## shape talk 0.06 0.06 0.10 0.10 0.30 0.20 0.18 0
## short tall 0.02 0.06 0.04 0.22 0.30 0.28 0.08 0
## talk school 0.00 0.04 0.02 0.08 0.10 0.28 0.48 0
## tell kids what to do 0.20 0.32 0.10 0.14 0.10 0.12 0.02 0
## tv conversation 0.04 0.00 0.02 0.10 0.30 0.40 0.14 0
## tv good for words 0.12 0.16 0.12 0.34 0.16 0.08 0.02 0
## words predict academic 0.14 0.10 0.10 0.14 0.18 0.28 0.06 0So \(\alpha = .71\) for these items, suggesting that they hang together relatively OK but that they still could be better. I think we could easily drop some items from the knowledge portion of the questions and do somewhat better here. In particular, read all words, no vocab bad in school, and tv conversation are actually negatively correlated with the scale as a whole.
4 Principal components analysis
Next, let’s look at a principal components analysis for the data. This approach tries to understand underlying orthogonal dimensions of variability.
4.1 All items
Explore using PCA to group items.
pcs <- princomp(x= select(wide.all, -workerid))
plot(pcs)
So there are many diffent reliable components here. Is this different if we normalize the likert to 0-1? No. (Not shown).
biplot(pcs)
This is pretty hard to interpret. Let’s go to the individual parts of the test.
4.2 Attitude items
First the attitude items.
row.names(wide.attitudes) <- wide.attitudes$workerid
pcs <- princomp(x = select(wide.attitudes, -workerid))
plot(pcs)
These items actually have three diffferent pretty significant components. So that’s interesting - there’s substantial heterogeneity on this measure. I wonder if we need more items?
Let’s try to plot the items by hand for more detail.
pc.items <- data.frame(pc1 = pcs$loadings[,1],
pc2 = pcs$loadings[,2],
pc3 = pcs$loadings[,3],
short_sent = row.names(pcs$loadings)) %>%
left_join(labels)
qplot(pc1, pc2, col = category,
label = short_sent, hjust = 1,
geom = c("text","point"),
data = pc.items) +
scale_colour_solarized() +
xlim(-.5,.5)
I guess we can speculate about what’s going on here - seems like we have some sense that PC1 is kind of more behaviorist, about reward, surprise, impulses, beahvioral control, and PC2 is maybe a little more cognitivist? Very hard to tell with this small a set of items.
Let’s look at individuals as well as items.
subinfo <- d %>%
group_by(workerid) %>%
summarise(children = children[1])
pc.inds <- data.frame(pc1 = as.numeric(pcs$scores[,1]),
pc2 = as.numeric(pcs$scores[,2]),
pc3 = as.numeric(pcs$scores[,3]),
workerid = row.names(pcs$scores)) %>%
left_join(subinfo)
qplot(pc1, pc2, col = children,
hjust = 1,
data = pc.inds) +
scale_colour_solarized()
The parents are perhaps a little further right, maybe a little less behaviorist?
4.3 Knowledge items
I’m expecting less in the way of structure for these items, but let’s try it anyway.
wide.knowledge <- d %>%
filter(instrument == "knowledge") %>%
mutate(rating = ifelse(reverse_code == 1, 7 - rating, rating)) %>%
select(workerid, short_sent, rating) %>%
spread(short_sent, rating)
pcs <- princomp(x= select(wide.knowledge, -workerid))
plot(pcs)
Here there’s really one big principal component, consistent with the impression that the scale is fairly reliable and unidimensional.
pc.data <- data.frame(pc1 = pcs$loadings[,1],
pc2 = pcs$loadings[,2],
pc3 = pcs$loadings[,3],
short_sent = row.names(pcs$loadings)) %>%
left_join(labels)
qplot(pc1, pc2, col = category,
label = short_sent, hjust = 1,
geom = c("text","point"),
data = pc.data) +
xlim(-.7,.3) +
scale_colour_solarized()
It’s hard for me to interpret this, seems like the variation is on PC2 not PC1, but at least it’s clear that read all words and counting before math are pretty different from others in the scale.
5 Conclusions
Looks like we should:
- Prune some of the knowledge items that are inconsistent or confusing
- Dichotomize responding on these items so we can assess accuracy and deal effectively with reverse-coded items
- Add some attitudes items to get a better picture of beliefs, since this measure seems like it could be revealing some differences.