My old laptop’s hard drive died, but thankfully I recognized the signs well in advance to back up my data and I am considering in replacing the old drive with a solid state drive. But I have decided to buy a new laptop and I am enjoying the games it can run. Back home in Montreal for R & R, I have been playing ANNO 2070 and in every session I have been reminded that I played for over 2 hours to which I thought I had played for a mere half hour. On the subject of time perception, there are some researchers here and there who investigated time perception and videogame play, but there has been a group led by Simon Grondin (Université Laval) who specialize in time perception research and persisted in their investigation for videogames. This is one of their latest publications.
Children’s time estimation literature lacks of studies comparing prospective and retrospective time estimates of long lasting ecological tasks, i.e. tasks reflecting children’s daily activities. In the present study, children were asked to estimate prospectively or retrospectively how much time they played a video game or read a magazine. Regardless of the task, the results revealed that prospective time estimates were longer than the retrospective ones. Also, time estimates of the video game task were longer, less accurate and more variable than those of the reading task. The results are discussed in the light of the current literature about time estimation of long lasting ecological tasks.
As of this posting, I am currently en route to Columbus for another year as a PhD student. Where did my two weeks went?
The study’s literature review is mostly pointing out gaps in the time perception literature and that the study addressed these gaps. The first gap is the comparative validity between laboratory tasks and ecological ones. The authors noted that tasks used in laboratory settings don’t exactly have anything relatable in daily life, this is especially noted for children, and there are few studies involving everyday activities. Second, there are developmental aspects in time perception where children between 3 to 8 years of age become sensitive and that there were very few studies involving children. Third, most time perception studies involve time estimation between 100ms to few seconds, studies for longer durations are few.
Core concepts in the literature involve two time estimation paradigm: prospective and retrospective. Prospective time estimation is where the individual is told beforehand they will be asked to judge time after finishing a task. Retrospective time estimation is where the individual is not told beforehand and were asked to estimate time when they finished a task. The key difference between prospective and retrospective is the processes involved. Prospective would involve active cognitive or attentional processes because the individual knows that a critical task, time estimation, is involved and will pay attention to time. Retrospective time estimation relies on memory as the individual is not actively paying attention to time, therefore how the mind can tell time when asked as an afterthought, the information available would be stored in memory. That is the gist of these two concepts, there are more fine details, but I leave it to the curious ones to read Simon Grondin’s review of the literature.
Participants: 199 elementary students from Quebec City. Ages from 8 to 12 years old. 94 boys and 105 girls. The participants came from four different elementary schools and obviously speak French and are learning English (starting from 4th grade since I started that way too).
Time estimation: students were asked the likely minimum and maximum duration of their task, be it videogame or reading.
Demographics and manipulation checks: age, gender, etc. Students responded to two questions pertaining to their level of enjoyment using a 7-point Likert-scale and their level of competence using a 4-point Likert-scale.
Videogame used: The Sims. My first thought is the videogame has to be easy to master and accessible to everyone. My second thought is that the videogame might have to be on the same level, of what kind I can’t tell, as the other task: reading.
Book used: a French magazine called Astrapi that contains articles and educational comics. I used to read J’aime Lire as a kid (mostly the comics section).
The classes are randomly assigned to one of four conditions from two factors: prospective versus retrospective. Videogame vs. reading.Instructions were given to the children on how to play the Sims. In the prospective condition, the experimenter told the students before they start that they would have to estimate their time. The objective time the students spent in the activities is 14 minutes.
There are three ways for subjective time estimation. Read the article for the details. Ratio is a comparison between estimated time versus target time (i.e. 14 minutes). Absolute standardized error (ASE) is the accuracy for estimated time. Weber Fraction (WF) like index is the variability in time estimation.
I will be focusing on the videogame vs. reading part of the study.
The ANOVA analysis for ratio revealed two main effects, but no interaction effects. Those in the prospective condition showed greater overestimation of time. Those in the videogame condition, quite surprisingly, showed greater overestimation of time than those in the reading condition. The level of enjoyment was not a significant factor when it was taken into account. However, the level of competence was found to be significant factor as the level of competence is negatively correlated with the ratio variable. Greater level of competence is correlated to lower overestimation of time. This might explain why there is time overestimation as those in the videogame group reported lower levels of competence compared to the reading group.
The authors conducted further analysis whether there was a mediated relationship between time estimation, task and time estimation. As I was reading it for a few times, I must ask if there were other ways for a mediation analysis as they simply conducted two separate correlations (videogame and reading) between competence level and time estimation. Their correlational analyses revealed no mediation and concluded that the level of competence did not significant mediate between task and time estimation.
The results for ASE revealed a main effect where those in the videogame condition had greater ASE scores than those in the reading condition, thus less accurate time estimation. Further investigation revealed an interaction effect where those in the prospective and videogame condition showed greater ASE scores. When accounting for enjoyment and competence, both were not significantly related to ASE.
The results for WF index revealed two main effects. Those in the videogame condition had greater WF index scores than the reading condition and boys had larger WF index scores than girls. Thus, greater WF index scores mean greater variability in time estimation in the sample. When accounting for enjoyment and competence, enjoyment was significantly negatively correlated to WF index scores. However, competence was not correlate to WF index scores.
The take home message is they have filled some gaps in the time estimation literature and are encouraging others to do more is what my impression is. As for the overestimation of time in the videogame condition, the authors argued that it could be due to videogame’s known effect of sucking one’s time that the children were adjusting their time estimation to compensate for their apparent loss of time awareness. Despite the non-differences in the levels of competence and enjoyment, I think they should use more rigorous manipulation checks using measures based on self-determination theory, flow theory, presence, and videogame experience among others. A more recent study (Gable & Poole, 2012) found that approach motivation causes a shortening perception of time, this means that people who want to play videogames are likely to experience distorted time perception. On the opposite spectrum (O’Brien et al., 2011), doing boring tasks, but not fun tasks, and especially when you felt entitled makes time seem to stand still.
On a sidenote, perhaps they should investigate the possible relationship between the zeigarnik effect and time perception. When playing ANNO 2070, I felt that time slows down when I am working towards a goal, such as trying to establish a secure supply of iron and coal and thinking the task is simple, I did not realize how long it took until the computer told me that two hours has passed.
The authors proposed another explanation is that there is an “adaptation period” where starting a new game might be less pleasant and thus increase an overestimation of time. This seems quite plausible since levels of competence is negatively correlated to ratio scores.
The study’s limitations are that a single duration (i.e. 14 minutes) was used in the experiment. They found in one of their previous studies that different durations showed different time estimation variability. An ecological task would trade off something from a laboratory task, namely experimental control. The authors gave an example that it is possible that some children’s arousal was increased and thus it is possible that arousal may have affected their time perception. Another limitation that I am incidentally looking into is the cognitive demands of the videogame. In one of their previous study, they used Tetris and found time underestimation which makes the argument possible that The Sims may be less cognitively demanding and thus the children were overestimating time. There is some support from the self-control literature (Vohs & Schmeichel, 2003) where exerting self-control can distort time perception.