table(HRS$R1213CHGRETSATIND)                       
## 
##    0    1 
## 5912  728
#728 people had a decrease in retirment satisfaction

table(HRS$R1213CHGRETSATIND, HRS$R1213MENHTHCHIND) 
##    
##        0    1
##   0 5120  471
##   1  610   57
#top is satisfaction. left is retirement satisfaction. 57 people who had a decrease in retirment satisfaction also developed sadness/depression/ phsy. 
#471 people had a drop in retirement satisfaction with no change on mental health. 610 people did not decrease their satisfaction from very but 
#developed a mental health issue. 5120 people reported no change in either. 

table(HRS$R1213CHGRETSATIND, HRS$R1213ADLCHID)      
##    
##        0    1
##   0 5315  542
##   1  622  101
#101 people had a drop in their retirement satisfaction when they devloped 'some difficulty in their ADL' 542 had a drop in retirment satisfaction 
#without an increase in daily living. 622 developed some difficulty in their ADL while retirment satisfaction remained consistent.

table(HRS$R1213CHGRETSATIND, HRS$R1213HDOCDIGID)    
##    
##        0    1
##   0 5541  320
##   1  677   41
# 42 people developed a drop RETSAT with a development of a major health condition (stoke,cancer,lung). 340 people had a drop in satisfaction with no 
# major health increase. 676 people developed a major health condition but it did not affect their retirment satisfaction 

table(HRS$R1213CHGRETSATIND, HRS$R1213CHGSLFHTHID)
##    
##        0    1
##   0 5484  418
##   1  645   82
#82 people reported a drop in their retirment satisfaction and drop in self reported health 418 people had a change in retirment satisfaction 
#But not change in health. 645 people had a change in health but not a change in retirment satisfaction.

So I was a little surprised to see these results as I expected a stronger relationship. I thought there would be a higher association between people that develop/have a mental health problem and a drop in their retirement satisfaction. This leads back to the question: “All in all, would you say that your retirement has turned out to be very satisfying, moderately satisfying, or not at all satisfying?” To me the question on retirement satisfaction is asking for an entire picture. Just because some one had a health issue does not make all there previous years in retirement a unsatisfying. This is also a potential reason for why their is small change in retirement satisfaction with only 728 people doping their retirement satisfaction or about 10% of the total sample. However having these issues for the duration of the retirement will have negative effects on retirement satisfaction as the ordinal logistic regression shows for wave 4.

summary(glm(R1213CHGRETSATIND~ R1213MENHTHCHIND+R1213ADLCHID+R1213HDOCDIGID+HRS$R1213CHGSLFHTHID, data = HRS)) 
## 
## Call:
## glm(formula = R1213CHGRETSATIND ~ R1213MENHTHCHIND + R1213ADLCHID + 
##     R1213HDOCDIGID + HRS$R1213CHGSLFHTHID, data = HRS)
## 
## Deviance Residuals: 
##     Min       1Q   Median       3Q      Max  
## -0.1880  -0.1013  -0.1013  -0.1013   0.9123  
## 
## Coefficients:
##                       Estimate Std. Error t value Pr(>|t|)    
## (Intercept)           0.101281   0.004437  22.825  < 2e-16 ***
## R1213MENHTHCHIND     -0.004240   0.014224  -0.298  0.76566    
## R1213ADLCHID          0.041416   0.013845   2.991  0.00279 ** 
## R1213HDOCDIGID       -0.009313   0.017832  -0.522  0.60152    
## HRS$R1213CHGSLFHTHID  0.045320   0.016275   2.785  0.00537 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## (Dispersion parameter for gaussian family taken to be 0.09537892)
## 
##     Null deviance: 588.45  on 6155  degrees of freedom
## Residual deviance: 586.68  on 6151  degrees of freedom
##   (35897 observations deleted due to missingness)
## AIC: 3011
## 
## Number of Fisher Scoring iterations: 2
#Based on the regression model it shows that a drop in daily living along with a drop in self reported health are the most impactful

Here is the results of the binary logistic model. Here it is showing that developing a difficulty in ADL and a decrease in self reported health are significant as coefficients when predicting a decrease in retirement satisfaction. Mental health and major health diagnosis are not. One concern that I have with this analysis here is how small of a sample is being looked at here.