Written and maintained by CASRAI Editorial Board
Published
On this page
The Strengths and Difficulties Questionnaire (SDQ) is a 25-item behavioural screening questionnaire for children and young people aged 2 to 17, scored as five scales of five items each. Researchers use it as a brief, multi-informant measure of emotional and behavioural problems and of prosocial behaviour. It is a screening and research instrument, not a diagnostic test, and this page does not support self-diagnosis. It explains how the scores are built, how to read them, and how to report and compute them. The scoring rules below follow the UK scoring guide dated 20 June 2016 and the official SDQ site (sdqinfo.org), both read 2026-11-08.
What the SDQ measures
The SDQ has 25 items, each answered Not True, Somewhat True or Certainly True. The items fall into five scales of five items: emotional symptoms, conduct problems, hyperactivity/inattention, peer relationship problems and prosocial behaviour. The first four scales are problem scales; the fifth is a strengths scale and is deliberately kept out of the problem total. The original instrument is described in Goodman (1997) and its psychometric properties in Goodman (2001) (see Sources). For the general measurement ideas behind a multi-item scale, see psychometrics.
Scales and items: which item belongs where
| Scale | Items | Reverse-scored items |
|---|---|---|
| Emotional symptoms | 3, 8, 13, 16, 24 | none |
| Conduct problems | 5, 7, 12, 18, 22 | 7 (generally obedient) |
| Hyperactivity/inattention | 2, 10, 15, 21, 25 | 21 (thinks before acting), 25 (sees tasks through) |
| Peer relationship problems | 6, 11, 14, 19, 23 | 11 (has a good friend), 14 (generally liked) |
| Prosocial behaviour | 1, 4, 9, 17, 20 | none (higher means more prosocial) |
SDQ scoring: how the scores are calculated
Item scores
Somewhat True always scores 1. Not True and Certainly True score 0 and 2 respectively on most items, and 2 and 0 on the reverse-scored items in the table above. Each scale therefore ranges from 0 to 10 when all five items are answered.
Missing items: the pro-rata rule
If at least 3 of the 5 items in a scale were answered, the scale score is scaled up pro rata to five items. The UK guide’s own example is a score of 4 on 3 completed items, which becomes 4 × 5 / 3 = 6.67, rounded to 7. With fewer than 3 answered items the scale is treated as missing.
Total difficulties, internalising and externalising
Total difficulties is the sum of the emotional, conduct, hyperactivity and peer scales (20 items, range 0 to 40); the prosocial scale is excluded, and the total is counted as missing if any of its four component scales is missing. The alternative three-scale grouping adds emotional and peer scores into an internalising score (0 to 20) and conduct and hyperactivity scores into an externalising score (0 to 20). The UK guide notes that the two combined scales may be preferable in community samples, while the four separate scales may add more value in high-risk samples (Goodman & Goodman, 2009).
Impact supplement
Versions with the optional impact supplement add questions on overall distress and interference with daily life. Responses score Not at all = 0, Only a little = 0, A medium amount = 1 and A great deal = 2, and the items are summed (0 to 10 for parent and self-report, 0 to 6 for teacher report). The chronicity and burden-to-others questions are not part of the impact score, and a “no” to the first supplement question scores the impact as zero.
Worked example (illustrative parent report)
The responses below are an invented illustrative case, built only to show the arithmetic; they do not describe a real child. Raw item scores already have the reverse-scoring applied.
| Scale | Item scores | Calculation | Scale score |
|---|---|---|---|
| Emotional | 1, 0, 2, 0, 1 | 1 + 0 + 2 + 0 + 1 | 4 |
| Conduct | 1, 0 (item 7 Certainly True), 0, 0, 0 | 1 + 0 + 0 + 0 + 0 | 1 |
| Hyperactivity | 2, 1, 2, 2 (item 21 Not True), 1 (item 25 Somewhat True) | 2 + 1 + 2 + 2 + 1 | 8 |
| Peer (items 19 and 23 missing) | 1 (item 6), 1 (item 11 Somewhat True), 2 (item 14 Not True) | (1 + 1 + 2) × 5 / 3 = 6.67, rounded to 7 | 7 |
Total difficulties = 4 + 1 + 8 + 7 = 20. Internalising = 4 + 7 = 11 and externalising = 1 + 8 = 9. Without the pro-rata step the peer score would be 4 and the total 17, which falls in a different band (see below), so the missing-data rule changes the category and should be applied the same way for every participant.
Interpretation: 3-band and 4-band cut-points
Scores can be used as continuous variables, and the UK guide says this is safest for the 18+ version, because cut-points have not been validated for it. The original bands were set from a UK population survey so that about 80% of children scored normal, 10% borderline and 10% abnormal. The newer four-band system, from a larger UK community sample, splits the top group: about 80% close to average, 10% slightly raised, 5% high and 5% very high. The guide calls both systems only a rough-and-ready way of screening, with symptom and impact scores from several informants better than any single report.
| Total difficulties, ages 4-17 | 3-band (normal / borderline / abnormal) | 4-band (close to average / slightly raised / high / very high) |
|---|---|---|
| Parent report | 0-13 / 14-16 / 17-40 | 0-13 / 14-16 / 17-19 / 20-40 |
| Teacher report | 0-11 / 12-15 / 16-40 | 0-11 / 12-15 / 16-18 / 19-40 |
| Self-report | 0-15 / 16-19 / 20-40 | 0-14 / 15-17 / 18-19 / 20-40 |
Because the thresholds differ by informant, the same total can land in different bands for a parent and a teacher. Subscale cut-points also exist in the scoring guide; consult the table in the guide itself rather than reusing the total difficulties thresholds. In the worked example, a parent total of 20 is in the “very high” band, and a hand-computed 17 without pro-rata scaling would be “high”.
Versions: parent, teacher, self-report and preschool
According to the official site, the parent and teacher versions cover ages 4 to 17 and the self-report version is intended for roughly ages 11 to 17. A modified preschool version for ages 2 to 4 keeps 22 identical items, softens one item, and replaces two antisocial items with oppositionality items. Follow-up versions add two intervention questions and use a one-month reference period. The official scoring page also states that it discourages hand scoring of paper forms, recommends online scoring, and notes that scoring syntax for SPSS, SAS, Stata and R is available with its own licence. For questionnaire downloads and translations, check the SDQ website’s notes on downloading and translations.
Reliability and validity: what to check in your own sample
Goodman (2001) examined the SDQ’s psychometric properties in a large British sample and supported its five-scale structure. Because reliability is a property of scores in a sample, report Cronbach’s alpha for each scale you analyse rather than citing values from other populations; five-item scales tend to give modest alphas. For validity evidence, see construct validity and convergent and discriminant validity. When parent and teacher reports disagree, treat that as information about informants and settings, not as error to be averaged away; inter-rater reliability coefficients are a way to quantify agreement. Check score distributions for floor effects in community samples, where many children score near zero on the problem scales.
Assumptions and common mistakes
- Forgetting to reverse-score items 7, 11, 14, 21 and 25.
- Including the prosocial scale in total difficulties.
- Applying UK cut-points to a very different population without checking, or applying 4-17 cut-points to the 18+ version.
- Mixing informants: parent, teacher and self-report cut-points differ.
- Treating a band as a diagnosis. The bands are screening categories.
- Scaling up scales with fewer than 3 answered items.
How to report the SDQ (APA style)
State the version, informant, age range, scoring method and how missing items were handled, then give descriptives and reliability from your sample. Example wording (replace the bracketed values with your own results):
Emotional and behavioural difficulties were measured with the parent-report Strengths and Difficulties Questionnaire (SDQ; Goodman, 1997). Total difficulties scores (range 0-40) were calculated by summing the emotional, conduct, hyperactivity and peer problem scales, with scale scores prorated when at least three of five items were completed. Internal consistency in the present sample was acceptable (α = [value]). Higher scores indicate greater difficulties.
How to calculate SDQ scores in Excel, R or Python
These snippets use our own logic for the published scoring rules (items coded 0 = Not True, 1 = Somewhat True, 2 = Certainly True before reversing). They are not the official scoring code.
Excel
Reverse-score an item (items 7, 11, 14, 21, 25): =IF(A2="","",2-A2) Scale score with pro-rata (five item cells B2:F2): =IF(COUNT(B2:F2)>=3,ROUND(SUM(B2:F2)/COUNT(B2:F2)*5,0),"") Total difficulties: =IF(COUNT(G2:J2)=4,SUM(G2:J2),"")
R
rev <- c(7, 11, 14, 21, 25)
scales <- list(emo = c(3,8,13,16,24), con = c(5,7,12,18,22),
hyp = c(2,10,15,21,25), peer = c(6,11,14,19,23))
# d: data frame with columns q1..q25 coded 0/1/2 (NA if missing)
for (i in rev) d[[paste0("q", i)]] <- 2 - d[[paste0("q", i)]]
scale_score <- function(x) ifelse(sum(!is.na(x)) >= 3,
floor(mean(x, na.rm = TRUE) * 5 + 0.5), NA)
for (nm in names(scales)) d[[nm]] <- apply(d[paste0("q", scales[[nm]])], 1, scale_score)
d$total <- d$emo + d$con + d$hyp + d$peer
Python
REVERSE = {7, 11, 14, 21, 25}
SCALES = {"emo": [3, 8, 13, 16, 24], "con": [5, 7, 12, 18, 22],
"hyp": [2, 10, 15, 21, 25], "peer": [6, 11, 14, 19, 23]}
def scale_score(resp, items): # resp: {item_number: 0/1/2 or None}
vals = [(2 - resp[i]) if i in REVERSE else resp[i]
for i in items if resp.get(i) is not None]
if len(vals) < 3:
return None
return int(sum(vals) * 5 / len(vals) + 0.5)
def total_difficulties(resp):
s = [scale_score(resp, v) for v in SCALES.values()]
return None if None in s else sum(s)
Frequently asked questions
What is the SDQ (Strengths and Difficulties Questionnaire)?
It is a brief behavioural screening questionnaire with 25 items in five scales, completed by parents, teachers or young people, covering ages 2 to 17 according to the official SDQ site.
How is the SDQ scored and interpreted?
Items score 0, 1 or 2 (with five reverse-scored items), five-item scales range 0 to 10, and total difficulties (0 to 40) sums four scales. Interpretation uses continuous scores or published 3-band and 4-band cut-points that differ by informant.
Is the SDQ a diagnostic tool?
No. The scoring guide describes the bands as a rough-and-ready screening aid; a questionnaire score cannot establish a diagnosis.
Where do I find the SDQ manual, scoring guide or downloads?
The official site is sdqinfo.org, which hosts the questionnaires, scoring information and notes on downloading and translations.
How do I score the SDQ if some items are missing?
Scale scores are prorated if at least 3 of 5 items were answered; otherwise the scale is missing, and total difficulties is missing if any of its four scales is.
Sources
- Youth in Mind. Strengths and Difficulties Questionnaire (SDQ) information, sdqinfo.org/a0.html and scoring page sdqinfo.org/py/sdqinfo/c0.py, read 2026-11-08.
- Scoring the Strengths & Difficulties Questionnaire for age 4-17 or 18+, 20 June 2016 (UK scoring guide, copy hosted by Bedford Borough Council), read 2026-11-08.
- Goodman, R. (1997). The Strengths and Difficulties Questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38(5), 581-586.
- Goodman, R. (2001). Psychometric properties of the Strengths and Difficulties Questionnaire. Journal of the American Academy of Child & Adolescent Psychiatry, 40(11), 1337-1345.
- Goodman, A., & Goodman, R. (2009). Strengths and Difficulties Questionnaire as a dimensional measure of child mental health. Journal of the American Academy of Child & Adolescent Psychiatry, 48(4), 400-403.







