Ethnicity and Greenspace Purpose (Q6_1 to Q6

Meet new people

tab <- table(data$Ethnicity, data$Q6_1)
chi <- chisq.test(tab)
resid <- round(chi$stdres, 2)
res_df <- as.data.frame(as.table(resid))
colnames(res_df) <- c("Ethnicity", "Response", "Z")

res_df$Sig <- cut(abs(res_df$Z),
                  breaks = c(0, 1.96, 2.58, Inf),
                  labels = c("", "*", "**"))

res_df$Label <- paste0(res_df$Z, res_df$Sig)

ggplot(res_df, aes(x = Ethnicity, y = Response, fill = Z)) +
  geom_tile(color = "white") +
  geom_text(aes(label = Label), size = 3) +
  scale_fill_gradient2(low = "blue", mid = "white", high = "red", midpoint = 0) +
  labs(
    title = "Adjusted Standardised Residuals: To Meet new people",
    x = "Ethnicity",
    y = "Response (Likert or Frequency)"
  ) +
  theme_minimal()

Interpretation: The chi-square test examines whether different ethnic groups gave different responses to the question: “Indicate why you might use green spaces: To Meet New People”. Statistical outcome: χ² = 46.3, df = 30, p = 0.03 This result indicates a statistically significant difference in how ethnic groups report using greenspaces for meeting new people.

Black respondents were strongly over-represented in selecting “Strongly agree” (z = +4.78) and markedly under-represented in “Strongly disagree” (z = -4.82) and “Disagree” (z = -4.59). This pattern suggests that Black participants were significantly more likely than expected to use greenspaces as sites for social connection with new people.

White respondents, by contrast, were significantly more likely to select “Strongly disagree” (z = +4.55) and “Disagree” (z = +2.12), and notably under-represented in “Strongly agree” (z = -3.98) and “Agree” (z = -3.33). These findings suggest that White participants were considerably less inclined to use greenspaces to meet new people, possibly reflecting greater use of green spaces for solitude, physical activity, or time with familiar social groups rather than for social expansion.

Asian respondents were over-represented in “Disagree” (z = +2.73) and under-represented in “Agree” (z = -1.34), indicating ambivalence or reluctance toward socialising with new individuals in green spaces. Mixed ethnicity participants showed modest over-representation in “Neither agree nor disagree” (z = +2.03), suggesting a more neutral or variable orientation.

To Spend time with family

tab <- table(data$Ethnicity, data$Q6_2)
chi <- chisq.test(tab)
resid <- round(chi$stdres, 2)
res_df <- as.data.frame(as.table(resid))
colnames(res_df) <- c("Ethnicity", "Response", "Z")

res_df$Sig <- cut(abs(res_df$Z),
                  breaks = c(0, 1.96, 2.58, Inf),
                  labels = c("", "*", "**"))

res_df$Label <- paste0(res_df$Z, res_df$Sig)

ggplot(res_df, aes(x = Ethnicity, y = Response, fill = Z)) +
  geom_tile(color = "white") +
  geom_text(aes(label = Label), size = 3) +
  scale_fill_gradient2(low = "blue", mid = "white", high = "red", midpoint = 0) +
  labs(
    title = "Adjusted Standardised Residuals: To spend time with family",
    x = "Ethnicity",
    y = "Response (Likert or Frequency)"
  ) +
  theme_minimal()

Interpretation: The chi-square test examines whether different ethnic groups gave different responses to the question: “Indicate why you might use green spaces: To spend time with Family” The heatmap shows adjusted standardised residuals. Positive Z-scores (in red) suggest higher-than-expected responses; negative scores (in blue) suggest lower-than-expected. Asterisks (*) indicate significant deviations from expected patterns. Ethnic groups with “**” contribute the most to the observed difference. Statistical outcome: χ² = 58.2, df = 30, p = 0.002 This indicates a statistically significant association between ethnicity and greenspace use for family-related purposes.

Asian respondents were significantly more likely than expected to disagree that they used greenspaces to spend time with family (z = +2.01). This may suggest differing norms about leisure in greenspaces.

Respondents identifying as “Other” ethnicity were markedly over-represented in selecting “Neither agree nor disagree” (z = +2.89) and “Strongly disagree” (z = +0.93), further supporting the possibility that for some ethnic groups, greenspaces may not serve as primary venues for family interaction.

In contrast, White respondents were under-represented in the “Neither agree nor disagree” category (z = -1.18), indicating clearer attitudes—whether positive or negative—toward using greenspaces for family time. However, no group was significantly over-represented at the “Agree” or “Strongly agree” end of the scale, suggesting that while family use of greenspaces is widespread, it may not be equally important across ethnic lines.

To spend time with Friends

tab <- table(data$Ethnicity, data$Q6_3)
chi <- chisq.test(tab)
resid <- round(chi$stdres, 2)
res_df <- as.data.frame(as.table(resid))
colnames(res_df) <- c("Ethnicity", "Response", "Z")

res_df$Sig <- cut(abs(res_df$Z),
                  breaks = c(0, 1.96, 2.58, Inf),
                  labels = c("", "*", "**"))

res_df$Label <- paste0(res_df$Z, res_df$Sig)

ggplot(res_df, aes(x = Ethnicity, y = Response, fill = Z)) +
  geom_tile(color = "white") +
  geom_text(aes(label = Label), size = 3) +
  scale_fill_gradient2(low = "blue", mid = "white", high = "red", midpoint = 0) +
  labs(
    title = "Adjusted Standardised Residuals: To spend time with friends",
    x = "Ethnicity",
    y = "Response (Likert or Frequency)"
  ) +
  theme_minimal()

Interpretation: The chi-square test examines whether different ethnic groups gave different responses to the question: “Indicate why you might use green spaces: To spend time with Friends” Statistical outcome: χ² = 49.8, df = 30, p = 0.014 This indicates a statistically significant association between ethnicity and the likelihood of using greenspaces for spending time with friends.

Black respondents were notably over-represented in the “Strongly agree” category (z = +1.93) and under-represented in both “Disagree” (z = -1.50) and “Neither agree nor disagree” (z = -1.91). These near-threshold residuals suggest a relatively stronger inclination among Black respondents to use greenspaces for socialising with friends, although the differences are not extreme.

White respondents showed a positive residual in “Strongly disagree” (z = +1.58), indicating a slightly greater tendency to reject the idea of using greenspaces for this purpose. While not crossing the conventional significance threshold (z > 2), this finding may hint at differences in how friendship-based leisure time is spatially organised—perhaps reflecting preferences for private or commercial venues over public outdoor settings.

Asian respondents showed no standout over- or under-representation, though a mild negative residual was observed for “Strongly disagree” (z = -1.89), suggesting they were somewhat less likely than expected to completely reject the idea of using greenspaces with friends.

Overall, this pattern of residuals points to subtle but important ethnic variations in social use of green spaces. While not marked by highly polarised views, Black participants appear slightly more inclined toward greenspace socialisation with friends, and White respondents somewhat less so. These differences may reflect divergent experiences of greenspace accessibility, safety, cultural familiarity, or group-oriented versus individualised recreational preferences.

To reflect or have time alone

tab <- table(data$Ethnicity, data$Q6_4)
chi <- chisq.test(tab)
resid <- round(chi$stdres, 2)
res_df <- as.data.frame(as.table(resid))
colnames(res_df) <- c("Ethnicity", "Response", "Z")

res_df$Sig <- cut(abs(res_df$Z),
                  breaks = c(0, 1.96, 2.58, Inf),
                  labels = c("", "*", "**"))

res_df$Label <- paste0(res_df$Z, res_df$Sig)

ggplot(res_df, aes(x = Ethnicity, y = Response, fill = Z)) +
  geom_tile(color = "white") +
  geom_text(aes(label = Label), size = 3) +
  scale_fill_gradient2(low = "blue", mid = "white", high = "red", midpoint = 0) +
  labs(
    title = "Adjusted Standardised Residuals: To reflect or have time alone",
    x = "Ethnicity",
    y = "Response (Likert or Frequency)"
  ) +
  theme_minimal()

Interpretation: The chi-square test examines whether different ethnic groups gave different responses to the question: “Indicate why you might use green spaces: To relax and unwind” Statistical outcome: χ² = 35.4, df = 30, p = 0.23 (non-significant)

Although the overall association is not statistically significant, some patterns are notable. Asian respondents were over-represented in the “agree” (z = +2.45) and “neither agree nor disagree” (z = +1.56) categories, while being significantly under-represented in “strongly agree” (z = –2.69), suggesting ambivalence or moderate endorsement. White respondents were significantly under-represented in “agree” (z = –2.95) and “neither” (z = –2.52), implying that relaxation in greenspaces is less frequently cited by this group. Black respondents showed mild positive residuals in neutral categories, suggesting more generalised use patterns.

To enjoy entertainment

tab <- table(data$Ethnicity, data$Q6_5)
chi <- chisq.test(tab)
resid <- round(chi$stdres, 2)
res_df <- as.data.frame(as.table(resid))
colnames(res_df) <- c("Ethnicity", "Response", "Z")

res_df$Sig <- cut(abs(res_df$Z),
                  breaks = c(0, 1.96, 2.58, Inf),
                  labels = c("", "*", "**"))

res_df$Label <- paste0(res_df$Z, res_df$Sig)

ggplot(res_df, aes(x = Ethnicity, y = Response, fill = Z)) +
  geom_tile(color = "white") +
  geom_text(aes(label = Label), size = 3) +
  scale_fill_gradient2(low = "blue", mid = "white", high = "red", midpoint = 0) +
  labs(
    title = "Adjusted Standardised Residuals: To enjoy entertainment",
    x = "Ethnicity",
    y = "Response (Likert or Frequency)"
  ) +
  theme_minimal()

Interpretation: The chi-square test examines whether different ethnic groups gave different responses to the question: “Indicate why you might use green spaces: To enjoy entertainment” Statistical outcome: χ² = 54.9, df = 30, p = 0.004

White respondents were significantly over-represented in “neither agree nor disagree” (z = +2.23), while under-represented in “strongly agree” (z = –1.65), implying low use of greenspaces as entertainment venues. Mixed ethnicity respondents were significantly under-represented in the “agree” category (z = –2.23) and over-represented in “neither” (z = +1.64), pointing to indecision. Black respondents were under-represented across nearly all categories, particularly “strongly agree” (z = –2.37) and “neither” (z = –2.51), suggesting greenspaces may be perceived as less appropriate venues for entertainment.

To participate in community activities

tab <- table(data$Ethnicity, data$Q6_6)
chi <- chisq.test(tab)
resid <- round(chi$stdres, 2)
res_df <- as.data.frame(as.table(resid))
colnames(res_df) <- c("Ethnicity", "Response", "Z")

res_df$Sig <- cut(abs(res_df$Z),
                  breaks = c(0, 1.96, 2.58, Inf),
                  labels = c("", "*", "**"))

res_df$Label <- paste0(res_df$Z, res_df$Sig)

ggplot(res_df, aes(x = Ethnicity, y = Response, fill = Z)) +
  geom_tile(color = "white") +
  geom_text(aes(label = Label), size = 3) +
  scale_fill_gradient2(low = "blue", mid = "white", high = "red", midpoint = 0) +
  labs(
    title = "Adjusted Standardised Residuals: To participate in community activitie",
    x = "Ethnicity",
    y = "Response (Likert or Frequency)"
  ) +
  theme_minimal()

Black respondents were significantly over-represented in both “agree” (z = +3.80) and “strongly agree” (z = +4.82) responses, suggesting a particularly strong alignment with the use of greenspaces for community-based activities. Black respondents were also under-represented in the negative response categories, such as “strongly disagree” (z = –2.65) and “neither agree nor disagree” (z = –1.88), indicating that community engagement in greenspaces is not only prominent but confidently affirmed within this group.

Asian respondents were over-represented in “disagree” (z = +1.99) and under-represented in “agree” (z = –1.36), implying lower levels of participation in community-oriented greenspace use—possibly due to structural exclusions, culturally misaligned programming, or limited awareness of opportunities.

White respondents showed modest over-representation in “strongly disagree” (z = +1.84) and “neither agree nor disagree” (z = +0.81), alongside under-representation in “agree” (z = –1.79) and “strongly agree” (z = –1.83). This pattern suggests a more detached or individualised orientation toward greenspaces, where community-based uses may be less salient or less prioritised.

To support others (e.g., attending a friend’s performance)

tab <- table(data$Ethnicity, data$Q6_7)
chi <- chisq.test(tab)
resid <- round(chi$stdres, 2)
res_df <- as.data.frame(as.table(resid))
colnames(res_df) <- c("Ethnicity", "Response", "Z")

res_df$Sig <- cut(abs(res_df$Z),
                  breaks = c(0, 1.96, 2.58, Inf),
                  labels = c("", "*", "**"))

res_df$Label <- paste0(res_df$Z, res_df$Sig)

ggplot(res_df, aes(x = Ethnicity, y = Response, fill = Z)) +
  geom_tile(color = "white") +
  geom_text(aes(label = Label), size = 3) +
  scale_fill_gradient2(low = "blue", mid = "white", high = "red", midpoint = 0) +
  labs(
    title = "Adjusted Standardised Residuals: To support others (e.g., attending a friend's performance)",
    x = "Ethnicity",
    y = "Response (Likert or Frequency)"
  ) +
  theme_minimal()

Interpretation: The chi-square test examines whether different ethnic groups gave different responses to the question: “Indicate why you might use green spaces: To participate in coummunity activities” Q6_7: To support others (e.g., attending a friend’s performance) Statistical outcome: χ² = 42.1, df = 30, p = 0.064 Note: This result is marginally non-significant at the conventional 0.05 threshold but reveals important patterns warranting interpretation.

Black respondents were highly over-represented in the “agree” category (z = +4.71), indicating a strong tendency to use greenspaces for supportive social purposes such as attending friends’ events or performances. Conversely, Black respondents were under-represented in “neither agree nor disagree” (z = –2.65) and “strongly disagree” (z = –2.36), suggesting that indifference or outright rejection of this greenspace motivation is uncharacteristic of this group.

White respondents were significantly over-represented in “disagree” (z = +2.99) and “strongly disagree” (z = +1.46), while under-represented in “agree” (z = –2.45) and “strongly agree” (z = –2.39).

Asian respondents displayed over-representation in “neither agree nor disagree” (z = +1.38) and “disagree” (z = +1.06), and under-representation in “agree” (z = –1.30) and “strongly agree” (z = –1.97). This points to lower use of greenspaces for supportive or reciprocal social roles within this group, perhaps due to limited access, competing social norms, or the use of other venues for such support.

Mixed and Other ethnic groups exhibited relatively neutral distributions across categories, with z-scores mostly between –1.0 and +1.3. The only modest deviation was among Mixed respondents in the “agree” (z = –1.01) and “strongly agree” (z = –0.91) categories, hinting at slight under-engagement.