CFQ: 25 questions about the slips everyone has
Walking into a room and forgetting why. Reading a page and realizing none of it went in. Doubting whether you locked the door ten minutes after leaving. Broadbent and colleagues assembled 25 of these in 1982 and asked how often each one happened over the past six months. What they found was less obvious than the items suggest: the score does not track how well people perform on cognitive tasks.
One score, not four
The original paper tried to separate perceptual, memory and action failures and could not. Factor analyses since have produced anywhere from one factor to four, depending on the sample: Wallace and colleagues found four, Rast and colleagues found three (forgetfulness, distractibility, false triggering), and the abbreviated 12-item version found three of its own (clumsiness, retrieval, intention forgotten). A large confirmatory analysis (N = 839, ages 16 to 85, 62% women, mean age 44.12, SD 19.54) settled on a single factor and advised against using subscales at all. Treat the total as one measure of how often your attention slips.
Answers run 0 (never) to 4 (very often), nothing is reverse-scored, and the total runs 0 to 100. It takes five to eight minutes.
Reliability
| Measure | Value |
|---|---|
| Internal consistency, with lie items | alpha 0.79 (n = 98) |
| Internal consistency, lie items removed | alpha 0.89 |
| Retest, 21 weeks | r = 0.824 (n = 57) |
| Retest, 65 weeks | r = 0.803 (n = 32) |
| Retest, 16 months, nursing students | r = 0.54 (n = 73) |
| Retest, 1 to 2 years, managers | tau = 0.621 (n = 114) |
| Retest, 1 to 2 years, spouses | tau = 0.563 (n = 42) |
Stability across a year or more is what makes the CFQ look like a trait. Broadbent's team had expected the score to move with stress, and mostly it did not.
What it relates to
| Measure | Correlation |
|---|---|
| Other cognitive failure scales | 0.57 to 0.62 |
| Trait anxiety | 0.311 |
| External locus of control | 0.350 |
| Neuroticism (EPQ N) | 0.277 |
| Social desirability (EPQ L) | -0.177 |
| Vocabulary intelligence (Mill Hill) | -0.157 |
| Objective neuropsychological tasks | no relationship |
The weak correlations carry the story. Higher vocabulary scores went with slightly fewer reported failures, and the relationship with measured performance was absent. People who report many failures are not necessarily performing worse. They may notice more, or they may be more distressed. Scores track depression, anxiety, attention problems, hyperactivity and impulsivity, which is why the CFQ is best described as a measure of perceived cognitive difficulty rather than ability.
Stress changes what the score means. In high-pressure settings a high CFQ predicted more symptoms later (tau = 0.462). In low-pressure settings the same score predicted nothing (tau = 0.087). Spouses agree with the self-report at tau = 0.315 to 0.360, so some of it is visible from outside.
Reading a score
| Total | Reading |
|---|---|
| 0-25 | Fewer slips than most people report |
| 26-50 | Typical range |
| 51-75 | More slips than average |
| 76-100 | Frequent slips |
Group averages from the original samples, for scale: car factory workers 35.02 (SD 11.52) and technicians 36.65 (SD 9.41), both on the 0 to 4 format, and nursing students 52.48 (SD 14.52). The original paper also tested groups on 1 to 4 and 1 to 6 formats, and those means cannot be compared with these, which is a useful reminder that a CFQ number only means something next to the format and sample it came from.
Versions and limits
Two short forms exist: the 12-item abbreviated version (Wilhelm and colleagues, 2010) and CFQ 2.0 (Goodhew and Edwards, 2024), which cuts the list to 15 items and about three minutes. No published Chinese validation exists, so the Chinese items on this site were translated from the original list rather than adapted from a validated version.
A high total is not a diagnosis of anything. It says you lose track of things more often than most people do, which can come from stress, poor sleep, distraction or low mood, and it is worth mentioning to a clinician if it is affecting your work. This page is not a diagnosis and does not replace one.