Table 3 Summary of studies reporting linguistic analyses
Study characteristics | Setting, task and assessments | Measures and results |
|---|---|---|
De Moor (2002) Design: RCT N = 42 Population: Renal cell carcinoma Gender: 85.7% male Mean age: 56.4 Ethnicity: Not reported | Setting: Lab-based Task: EW: Cancer-related emotions (n = 21); Control: Health behaviours (n = 21) Four sessions for 20 min over four weeks Assessments: Immediately post-intervention and at four, six, eight, and 10 weeks | Linguistic analyses: conducted using the LIWC Results: EW and control groups differed in the words they used for 24 of the 32 categories suggesting emotional and cognitive processing and expression of their cancer experience |
Imrie & Troop (2012) Design: Non-randomised experiment N = 6 Population: Secondary cancer or life-limiting illness Gender: 61.5% female Mean age: 67.5 (SD = 14.9; range = 38–86) Ethnicity: Not reported | Setting: Day Hospice Task: EW: Difficult experience from the previous week followed by expressing compassion for the self in the entry (n = 3); Control: Difficult experience from the previous week (n = 3) Three sessions for 20 min over three weeks Assessments: Baseline and one-week post-intervention | Linguistic analyses: conducted using the LIWC Results: Both groups reduced the number of negative words they used between baseline and follow-up (F1,18 = 6.97, p < 0.02) but compared to the control group, the expressive writers increased the number of causal words used over time (F1,18 = 8.36, p < 0.01) |
Laccetti (2007) Design: Secondary analysis of EW entries from RCT N = 68 Population: Metastatic breast cancer Gender: All females Mean age: 51 (range = 36–78) Ethnicity: 94% White, 5% Native American, 1% Other | Setting: Outpatient clinics Task: EW: Four sessions for 20–30 min over four consecutive days about experiences, thoughts and feelings related to not fully recovering from cancer and facing death, and traumatic and upsetting experiences in life that may or may not relate to cancer Assessments: Within one week of study entry and three months post-intervention | Linguistic analyses: conducted using the LIWC Measures: FACT-B assessed quality of life Results: Expressive writers who used more positive emotion words reported higher scores on emotional well-being (β = 1.87 [95% CI 0.33, 3.42], p = 0.02) and concerns related to their breast cancer of FACT-B (β = 1.75 [95% CI 0.17, 3.33] p = 0.03) three months post-intervention compared to those who used more negative emotion words |
Mosher (2012) Design: RCT N = 86 Population: Advanced breast cancer Gender: All female Mean age: EW: 57.4 (SD = 12.5); Control: 58.5 (SD = 11.7) Ethnicity: 81.4% White, 7% African American, 5.8% Hispanic, 5.8% Other | Setting: Home based Task: EW: Cancer-related emotions (n = 44); Control: Previous day’s activities (n = 42) Four sessions for 20 min over four-seven weeks Assessments: Eight weeks post-intervention | Linguistic analyses: conducted using the LIWC Results: EW group used a higher proportion of positive (η2p = 0.13, p < 0.001) and negative (η2p = 0.46, p < 0.001) words compared to the control group |