Participants
The only inclusion criteria were 18 ≥ years old and English-speaking. The sample consisted of 54 participants (M age = 26.9, SD = 7.5, range 19–60; 66.7% female). They were recruited through advertisements on social media and snowballing. As their highest educational attainment, the majority reported having a bachelor (64.27%) or postgraduate degree (13.2%), and the remaining had a diploma or certificate (9.4%) or finished high school (13.2%). The majority identified as Caucasian/White European (45.3%), and the remaining as Asian (26.4%), Arab or Middle Eastern (15.1%), African (3.8%), or “Other ethnicity” (9.4%). Most reported being in paid employment (60.8%), and just over half were studying (56.6%).
An a priori power analysis was conducted using G*Power 3.1.9.2 (Faul et al.,
2007). Based on the smallest observed effect sizes from previous work using this guided future thinking activities (Hallford et al.,
2020b), namely for perceived control and behavioural intention, we powered the study to detect at least small to moderate-sized within-group effects (
dz = 0.40) using paired samples
t-tests, with a power of 0.80 and alpha level of 0.05. To detect this effect, we required a total sample size of 50. We oversampled to account for some dropout, recruiting a total of 54 participants. Two of these participants completed only baseline and content details condition. To utilise all collected data, the missing responses for the spatial condition were imputed using the expectation maximisation procedure. Sensitivity analysis indicated removing these cases did not have any substantive effects on the results.
Materials
Future thinking simulations
As part of the online survey at baseline, participants were then asked to select five, plausible, self-relevant future events that they were not already planning and that would be positive experiences for them. Participants were instructed they could nominate any type of event or activity and were given a range of categories as prompts: work/school/study, conversation/socializing, errands/chores, a hobby/interest, physical activity/sport, TV/internet/games/media, and ‘other’. They were instructed that the events were to occur within the space of a day and in a specific place, and some examples were given (e.g., going shopping at a local shopping mall, going for lunch with a friend at a particular restaurant). They were then asked to provide a brief description of the event. To account for temporal distance, the participants were asked to imagine the first event as occurring in the next week, the second in the next month, the third the next year, the fourth the next 5 years, and the fifth in the next 10 years. To assess the validity of this manipulation, after the event description they were asked how far into the future the event was imagined as taking place using a scale from 1 (the next 24 h) to 6 (the next 10 years). A repeated measures ANOVA comparing the mean of each of the five events on this temporal distance scale showed a clear main effect, F(4, 212) = 93.9, p < 0.001, \(\eta_{{\text{p}}}^{2}\) = 0.63. Inspection of the means and follow up t-tests showed that each event corresponded approximately with the intended temporal distance, with significantly higher temporal distance for each subsequent future event (all p < 0.001): 2.0 (SD = 0.8), 2.7 (SD = 0.8), 3.3 (SD = 0.9), 4.2 (SD = 1.2), 5.1 (SD = 1.5).
To replicate methodology from Rubin et al. (
2019), the same measures of content details, spatial details, and pre-experiencing were used, and rated using self-report response scales from 1 (not at all) to 7 (definitely). The content details items were, “When thinking about this future event, I can identify or name the setting where it occurs, although I might not be able to describe it clearly”, and “As I think about this future event, I can identify the actions, objects, and/or people that are involved in it, though I may not be able to say clearly where they are in relation to each other”. These were averaged together across the future events and showed acceptable internal reliability (
α = 0.73). The spatial details items were, “When thinking of this future event, I imagined a scene in which the elements of the setting were located relative to each other in space” and “When thinking about this future event, I can describe where the actions, objects, and/or people are located in my imagination”. These were also averaged, and showed acceptable internal reliability (
α = 0.80). The pre-experiencing items were, “While thinking about this future event, it is as if I am pre-living the event”, “While thinking of this future event, it is as if I am mentally traveling to the time and place of the occurrence”, and “While thinking of this future event, it is as if I am experiencing the feelings, emotions, and/or atmosphere that I would feel when it happens”. These were averaged across events, with acceptable internal reliability (
α = 0.92).
The following items all used a self-report responses scale from 1 (not at all) to 9 (very much so) and were averaged across the five future events. The item, “How vivid and detailed is your thought of doing this activity?”, was used to rate overall detail and vividness (α = 0.74), and the item, “How much did you find yourself thinking in pictures/mental pictures about this?” to measure mental imagery (α = 0.61). To rate anticipated (expected pleasure from the event) and anticipatory pleasure (pleasure experienced thinking about the event), respectively, the following items were used: “How pleasant/enjoyable do you think it will be to do this activity?” (α = 0.66), and ‘How pleasant/enjoyable is it to just think about doing this activity?’ (α = 0.53). For perceived control the following item was used: ‘How much control do you think you would have over that activity occurring? As in, how easy do you think it would be to do?’ (α = 0.69). Lastly, to assess the intention to engage in the future events, participants were asked “How likely is it that you will do this activity?” (α = 0.69).
The following measures of depressive symptoms and trait anticipatory pleasure were administered at baseline in order to help describe the sample.
Depression, anxiety, and stress scale (DASS; Lovibond & Lovibond, 1995)
Depressive symptoms were assessed at baseline using the depression subscale of the DASS 21-item short-form. The DASS depression subscale assesses core features of depression (e.g., low mood, loss of interest, self-worth, and motivation) 7-item self-report items, and has previously shown excellent psychometric properties (Antony et al.,
1998). The items are rated on a scale from 0 (did not apply to me ever) to 4 (Applied to me very much, or most of the time). Reponses are then summed, with higher scores indicating higher severity of symptoms. Internal reliability in the current sample was good (α = 0.87).
Temporal experience of pleasure scale (TEPS; Gard et al., 2006)
To assess trait levels of anticipatory pleasure, the anticipatory pleasure subscale of the TEPS was used. This scale uses 10 self-report items to assess the general tendency to think about positive future events and experience anticipatory pleasure. Responses were given on a 6-point scale from 1 (Very falSE for me) to 6 (Very true for me), and items were summed with higher scores indicating a stronger tendency to anticipate pleasure. Internal reliability in the current sample was acceptable (α = 0.68).
Future thinking simulation protocol
The future thinking protocol was adapted from a protocol developed in our lab and used in two previous experiments to elicit detail in episodic future thinking (Hallford et al.,
2020b). In this protocol, five events are generated by participants prior to any guided simulations, and then rated for various dimensions. This is used as the baseline condition. The same events are then used in the guided future thinking conditions. The participants were instructed to think of unique events that they would be personally involved in, or a unique instance of a type of event. This helped to ensure that participants used episodic future thinking, rather than just imagining event categories, referred elsewhere as personal semantics (Renoult et al.,
2012). Participants were given general instructions to imagine events from a first-person perspective, as if they were actually happening, and to use mental imagery. They were asked to think only about the particular event that that was being discussed, taking place in a specific location and time, and not about experiences occurring before or after this. They were first given 30s to describe the highlight of each event and which positive emotions they would be likely to feel during that event. Then participants were questioned about the event for two minutes, in a way that was dependent on the condition: (1) about the contents of the scene without describing the environment (e.g., the people, the sequences of events, objects; e.g., “PleaSE describe some objects that appear in this event”. “Can you tell me about the order in which things will happen?”) or (2) about the spatial and environmental details (e.g., the size and shape of room or location, features of the physical space, how far apart objects in the location were from each other; e.g., “Can you tell me some details about the physical dimensions of the place in which this would happen?”, “Can you tell me about the dimensions of the place?—height, width, length?”). The full protocol is provided in the Supplementary Materials. Prompts were used flexibly dependent on what details the participants gave. A practice trial was provided at the start of each condition in which participants simulated doing a neutral task of mailing a letter to their local post office. Experimenters were trained on the simulation protocol over a series of sessions, and piloting was conducted with two participants prior to recruiting the main sample.
Data analysis strategy
Mean scores and standard deviations were calculated for all study variables. Pearson correlation coefficients were used to test zero-order associations between the variables at baseline. For Aim 1, to assess the unique contribution of content and spatial details with the dependent variables, a series of multiple regressions were conducted. The variables of temporal distance, age, and gender were also entered as predictors, as per Rubin et al. (
2019). For Aim 2, to assess for changes in the dependent variables over the future thinking conditions, repeated measures ANOVAs were conducted, with planned paired samples
t-tests in accordance with the hypotheses, to compare differences between conditions where an omnibus effect was found. Cohen’s
dz was used to indicate the magnitude of effect between conditions.
In order of frequency, future events relating to work/school/study were most common (21.9%) along with ‘other activities not listed above’ (21.9%). Following these were conversation/socialising (17.8%), tv/internet/games/media (11.5%), physical activity (10.7%), eating/drinking (7%), a hobby (not physical activity; 4.8%), and errand/chores (4.4%). Full descriptive statistics for the variables are found in Table
1, and correlations between baseline variables in Table
2. Regarding the correlations, depressive symptoms were correlated with lower perceived control. Trait anticipatory pleasure (TEPS-A) was correlated with higher spatial and content details, pre-experiencing, and perceived control. All detail-related variables correlated with one another and with pre-experiencing and mental imagery. Spatial details were correlated with anticipatory pleasure and overall detail/vividness, but not anticipated pleasure, and content details were only related to overall detail/vividness. Perceived control correlated with all detail variables, but not pre-experiencing or mental imagery. Behavioural intention correlated with perceived control, but no other variables.
Table 1
Means and standard deviations of study variables in each condition
DASS depression | | – | – |
TEPS-A | | – | – |
Spatial details | 5.0 (1.0) | 5.6 (0.9) | 5.9 (0.7) |
Content details | 5.2 (0.9) | 5.7 (0.9) | 5.7 (0.8) |
Overall detail/vividness | 6.4 (1.4) | 7.0 (1.2) | 7.3 (0.9) |
Pre-experiencing | 4.5 (1.2) | 5.1 (1.1) | 5.2 (1.1) |
Mental Imagery | 6.6 (1.3) | 7.3 (1.1) | 7.6 (1.0) |
Anticipated pleasure | 7.7 (1.0) | 7.5 (0.8) | 7.5 (0.9) |
Anticipatory pleasure | 7.1 (1.1) | 6.9 (1.0) | 6.9 (1.1) |
Perceived control | 5.9 (1.5) | 6.3 (1.1) | 6.8 (1.0) |
Behavioural intention | 7.1 (1.1) | 7.0 (1.0) | 7.1 (0.9) |
Table 2
Zero-order correlations between study variables at baseline (N = 54)
1.DASS depression | – | | | | | | | | | | |
2. TEPS-A | – 0.09 | – | | | | | | | | | |
3. Spatial details | – 0.05 | 0.38** | – | | | | | | | | |
4. Content details | – 0.24† | 0.34* | 0.62*** | – | | | | | | | |
5. Overall detail/vividness | – 0.13 | 0.23† | 0.59*** | 0.38** | – | | | | | | |
6. Mental Imagery | – 0.11 | 0.21 | 0.48*** | 0.31* | 0.77*** | – | | | | | |
7. Pre-experiencing | – 0.04 | 0.39** | 0.59*** | 0.39** | 0.65*** | 0.55*** | – | | | | |
8. Anticipated pleasure | – 0.03 | 0.01 | 0.16 | – 0.07 | 0.46*** | 0.34* | 0.13 | – | | | |
9. Anticipatory pleasure | 0.08 | 0.22 | 0.34* | 0.13 | 0.62*** | 0.58*** | 0.47*** | 0.71*** | – | | |
10. Perceived control | – 0.31* | 0.32* | 0.37** | 0.38** | 0.32* | 0.22 | 0.23† | 0.16 | 0.22 | – | |
11. Behavioural intention | – 0.16 | 0.09 | 0.18 | 0.21 | 0.09 | 0.01 | 0.01 | – 0.18 | – 0.10 | 0.50*** | – |
Aim 2: manipulation checks
A 3 (condition: baseline, content, spatial) × 2 (detail type: content, spatial) ANOVA indicated a main effect for condition, F(2, 106) = 22.2, p < 0.001, \(\eta_{{\text{p}}}^{2}\) = 0.22, no main effect for detail type, F(1, 53) = 1.3, p = 0.245, \(\eta_{{\text{p}}}^{2}\) = 0.02, and a condition by detail type interaction effect, F(2, 106) = 4.6, p = 0.011, \(\eta_{{\text{p}}}^{2}\) = 0.81. Follow-up tests for content details indicated participants reported higher detail in the content condition relative to baseline (M difference = 0.41, 95% CI 0.16, 0.67; t(53) = 3.3, p = 0.002, dz = 0.45), and in the spatial condition relative to baseline (M difference = 0.48, 95% CI 0.21, 0.75; t(53) = 3.5, p = 0.001, dz = 0.49), but there was no significant difference between content and spatial conditions (M difference = 0.06, 95% CI − 0.10, 0.23; t(53) = 0.8, p = 0.426, dz = 0.11). Follow-up tests for spatial details indicated participants reported higher detail in the content condition relative to baseline (M difference = 0.60, 95% CI 0.32, 0.89; t(53) = 4.2, p < 0.001, dz = 0.57), and in the spatial condition relative to baseline (M difference = 0.87, 95% CI 0.60, 1.13; t(53) = 6.6, p < 0.001, dz = 0.90). Spatial details were reported as significantly higher in the spatial condition relative to the content condition (M difference = 0.26, 95% CI − 0.04, 0.48; t(53) = 2.3, p = 0.021, dz = 0.33). In summary, content details increased in both conditions from baseline to a similar degree, whereas spatial details increased in both conditions and were higher in the spatial condition relative to the content condition.
Aim 2: repeated measures analyses
Firstly, effects of the guided future thinking conditions were assessed on overall detail/vividness and mental imagery. A repeated measures ANOVA for detail/vividness showed a main effect for condition, F(2, 106) = 18.1, p < 0.001, \(\eta_{{\text{p}}}^{2}\) = 0.25. Participants reported higher detail/vividness in the content condition relative to baseline (M difference = 0.61, 95% CI 0.25, 0.98; t(53) = 3.3, p = 0.001, dz = 0.45), and in the spatial condition relative to baseline (M difference = 0.93, 95% CI 0.63, 1.23; t(53) = 6.2, p < 0.001, dz = 0.85), and overall detail/vividness was significantly higher in the spatial condition relative to the content condition (M difference = 0.31, 95% CI 0.03, 0.59; t(53) = 2.2, p = 0.028, dz = 0.30).
A main effect was also found for mental imagery, F(2, 62) = 18.9, p < 0.001, \(\eta_{{\text{p}}}^{2}\) = 0.26. More use of mental imagery was reported in the content condition relative to baseline (M difference = 0.71, 95% CI 0.37, 1.05; t(53) = 4.1, p < 0.001, dz = 0.56), and in the spatial condition relative to baseline (M difference = 0.96, 95% CI 0.60, 1.32; t(53) = 5.3, p < 0.001, dz = 0.72), but the trend for a difference between content and spatial conditions did not reach statistical significance (M difference = 0.25, 95% CI 0.01, 0.52; t(53) = 1.7, p = 0.068, dz = 0.25).
A main effect was found for pre-experiencing, F(2, 106) = 15.5, p < 0.001, \(\eta_{{\text{p}}}^{2}\) = 0.22. Participants reported stronger pre-experiencing in the content condition relative to baseline (M difference = 0.58, 95% CI 0.28, 0.88; t(53) = 3.9, p < 0.001, dz = 0.53), and in the spatial condition relative to baseline (M difference = 0.66, 95% CI 0.39, 0.94; t(53) = 4.8, p < 0.001, dz = 0.66), but the difference between content and spatial conditions was not significant (M difference = 0.08, 95% CI − 0.11, 0.28; t(53) = 0.8, p = 0.419, dz = 0.10).
Next, the anticipated and anticipatory pleasure items were assessed for change over the conditions. A main effect was found for anticipated pleasure, F(2, 106) = 3.2, p = 0.042, \(\eta_{{\text{p}}}^{2}\) = 0.05. Participants reported lower anticipated pleasure in the content condition relative to baseline (M difference = 0.24, 95% CI − 0.01, − 0.47; t(53) = 2.0, p = 0.041, dz = 0.24), but no significant difference in the spatial condition relative to baseline (M difference = 0.20, 95% CI − 0.00, − 0.41; t(53) = 1.9, p = 0.058, dz = 0.26). The difference between content and spatial conditions was not significant (M difference = 0.03, 95% CI − 0.08, 0.20; t(53) = 0.4, p = 0.649, dz = 0.00). No main effect for anticipatory pleasure, F(2, 106) = 2.2, p = 0.111, \(\eta_{{\text{p}}}^{2}\) = 0.04, indicating conditions did not significantly differ.
A main effect was found for perceived control, F(2, 106) = 10.6, p < 0.001, \(\eta_{{\text{p}}}^{2}\)0.16. Participants reported higher perceived control in the content condition relative to baseline (M difference = 0.46, 95% CI 0.01, 0.91; t(53) = 2.0, p = 0.045, dz = 0.31), and in the spatial condition relative to baseline (M difference = 0.88, 95% CI 0.51, 1.2; t(53) = 4.8, p < 0.001, dz = 0.70), and perceived control was significantly higher in the spatial condition relative to the baseline condition (M difference = 0.42, 95% CI 0.10, 0.74; t(53) = 2.6, p = 0.011, dz = 0.36). No main effect for condition was found for behavioural intention, F(2, 106) = 0.4, p = 0.631, \(\eta_{{\text{p}}}^{2}\) = 0.00, indicating conditions did not significantly differ.
As requested by a reviewer, to assess if there were repetition suppression effects or whether participants improved purely as a function of repeated guided simulation, regardless of condition, t-tests were conducted across the three conditions in the order they were presented for participants. Some condition order information was not available due to a clerical error, leaving n = 39 for these analyses. The results did not show evidence of either of these effects, with guided interview conditions both being higher on content, spatial, overall detail/vividness, and perceived control than baseline (all p’s < 0.05), and no difference between the first and second guided interviews (all p’s > 0.05). Consistent with the results above, for anticipated and anticipatory pleasure and behavioural intention, the guided interview conditions did not differ from baseline, nor the first and seconded guided interviews differ from each other (all p’s > 0.05).