Memo: Do regular transit riders differ in CA?

Research memorandum — observational PRCA contrast

Author

Jack J. Burleson

Published

Invalid Date

Research question: Do individuals who take public transportation regularly have communication-apprehension (CA) scores that differ from non-regular riders in the matched cohort?

Formal write-up: docs/secondary_rq_transit_ca.md

1 Answer, Response, + Summary of Results

Using the Prolific↔︎Qualtrics matched cohort (File A + File B stacked joined to File C; 252 matched rows; analytic n = 241 with complete PRCA items), we compared ground-truth group and interpersonal communication apprehension (CA; range 6–30; (McCroskey, 1970)) for regular public-transit riders vs everyone else. Regular transit = Q26 ∈ {4–8 days a month, 8 or more days a month} (weekly-or-more). Primary test: Welch t; effect sizes Cohen’s d / Hedges’ g; ****bootstrap 95% CIs**** (ca-personas transit-ca --join inner --seed 42).

Short answer: Yes. Regular riders report lower CA on both subscales (see Figure 1). Group CA shows the clearer gap (Δ = −2.72, d = −0.46, p = .0003); interpersonal CA is smaller but still significant (Δ = −1.73, d = −0.30, p = .020).

Figure 1.. Mean CA by `Q26` ridership and regular vs not-regular

Key numbers (regular n = 101 vs not-regular n = 140; Table Table 1):

Table 1.. Group and interpersonal CA by transit ridership status.

Subscale	Regular M	Not-regular M	Δ	Welch p	Cohen’s d
Group CA	13.04	15.76	−2.72	0.0003	−0.46
Interpersonal CA	13.31	15.04	−1.73	0.020	−0.30

Mean CA also falls as Q26 intensity rises (“Never” group CA M = 17.38; “8 or more days/month” M = 12.84).

Conclusion. Weekly+ public-transit use is associated with lower PRCA scores in this sample — especially group CA (Δ = −2.72, d = −0.46). That group difference is consistent with the companion CA→transit Random Forest (AUC = 0.590; ca_scores_predict_transit.qmd).

Sources: notebooks/secondary_rq_transit_ca.ipynb · src/ca_personas/transit_ca.py · write-up docs/secondary_rq_transit_ca.md · github.com/Exios66/psych755-jjb

2 What questions or uncertainties remain?

Is lower CA a cause of transit use, a consequence of routine shared-vehicle exposure, or a correlate of urbanicity / car access / employment? Same-wave data cannot decide. Whether the CA gap survives after accounting for ride-share exposure is tested in the follow-up residual_ca_after_rideshare.qmd (CA adds only +0.021 AUC over Q28; the within-Q28-strata CA gap is heterogeneous).

3 What other features may also well-predict regular public transit use?

Geography recovers AUC = 0.551 (geo_predicts_transit.qmd); CA scores jointly recover AUC = 0.590 (ca_scores_predict_transit.qmd). Head-to-head mobility follow-ups report car access AUC = 0.607, employment AUC = 0.528, and Q28 AUC = 0.762 (transit_covariate_followups.qmd; q27_q28_predict_transit.qmd).

References

McCroskey, J. C. (1970). Measures of communication-bound anxiety. Speech Monographs, 37(4), 269–277. https://doi.org/10.1080/03637757009375677

--- title: "Memo: Do regular transit riders differ in CA?" subtitle: "Research memorandum — observational PRCA contrast" author: Jack J. Burleson date: 2026–07-25 bibliography: ../references.bib --- **Research question:** Do individuals who take public transportation regularly have communication-apprehension (CA) scores that differ from non-regular riders in the matched cohort? **Formal write-up:** [`docs/secondary_rq_transit_ca.md`](../docs/secondary_rq_transit_ca.md) --- ## Answer, Response, + Summary of Results Using the Prolific↔Qualtrics matched cohort (File A + File B stacked joined to File C; **252** matched rows; analytic **n = 241** with complete PRCA items), we compared ground-truth **group** and **interpersonal** communication apprehension (CA; range 6–30; [@mccroskey1970]) for **regular** public-transit riders vs everyone else. Regular transit = `Q26` ∈ {`4–8 days a month`, `8 or more days a month`} (weekly-or-more). Primary test: Welch *t*; effect sizes Cohen's *d* / Hedges' *g*; ****bootstrap 95% CIs**** (`ca-personas transit-ca --join inner --seed 42`). **Short answer:** Yes. Regular riders report **lower** CA on both subscales (see @fig-tc-transit-riders-ca). Group CA shows the clearer gap (Δ = **−2.72**, *d* = −0.46, *p* = .0003); interpersonal CA is smaller but still significant (Δ = **−1.73**, *d* = −0.30, *p* = .020). ![Mean CA by `Q26` ridership and regular vs not-regular](figures/transit_riders_ca_memo.png){#fig-tc-transit-riders-ca} **Key numbers (regular n = 101 vs not-regular n = 140; Table @tbl-tc-1):** | Subscale | Regular M | Not-regular M | Δ | Welch *p* | Cohen's *d* | |---|---:|---:|---:|---:|---:| | Group CA | 13.04 | 15.76 | **−2.72** | 0.0003 | −0.46 | | Interpersonal CA | 13.31 | 15.04 | **−1.73** | 0.020 | −0.30 | : Group and interpersonal CA by transit ridership status. {#tbl-tc-1} Mean CA also falls as `Q26` intensity rises (“Never” group CA *M* = **17.38**; “8 or more days/month” *M* = **12.84**). **Conclusion.** Weekly+ public-transit use is associated with lower PRCA scores in this sample — especially **group** CA (Δ = −2.72, *d* = −0.46). That group difference is consistent with the companion CA→transit **Random Forest** (AUC = **0.590**; [`ca_scores_predict_transit.qmd`](ca_scores_predict_transit.qmd)). *Sources:* `notebooks/secondary_rq_transit_ca.ipynb` · `src/ca_personas/transit_ca.py` · write-up `docs/secondary_rq_transit_ca.md` · [github.com/Exios66/psych755-jjb](https://github.com/Exios66/psych755-jjb) --- ## What questions or uncertainties remain? Is lower CA a cause of transit use, a consequence of routine shared-vehicle exposure, or a correlate of urbanicity / car access / employment? Same-wave data cannot decide. Whether the CA gap survives after accounting for ride-share exposure is tested in the follow-up [`residual_ca_after_rideshare.qmd`](residual_ca_after_rideshare.qmd) (CA adds only +0.021 AUC over `Q28`; the within-`Q28`-strata CA gap is heterogeneous). ## What other features may also well-predict regular public transit use? Geography recovers AUC = **0.551** ([`geo_predicts_transit.qmd`](geo_predicts_transit.qmd)); CA scores jointly recover AUC = **0.590** ([`ca_scores_predict_transit.qmd`](ca_scores_predict_transit.qmd)). Head-to-head mobility follow-ups report car access AUC = **0.607**, employment AUC = **0.528**, and `Q28` AUC = **0.762** ([`transit_covariate_followups.qmd`](transit_covariate_followups.qmd); [`q27_q28_predict_transit.qmd`](q27_q28_predict_transit.qmd)).