• Sociology: research methods

What is the difference between reliability and validity in research?

Written with AI (How we create content)Target level: Grade 11Last updated: Report an error
Quick answer

The difference between reliability and validity is the difference between consistent and correct. A measure is reliable when it gives the same result again, at another time or in another researcher's hands. It is valid when it measures what it claims to measure. A measure can be reliable without being valid, but a measure that is not reliable cannot be valid.

The topic explained

Two questions to ask of every measure

Every finding in sociology rests on a measure: a questionnaire, an interview guide, an observation sheet, an official statistic. Two questions decide whether it can be trusted, and the difference between reliability and validity is the difference between them. Reliability asks whether the measure is consistent: would it give the same result if the study were repeated, at another time or by another researcher? Validity asks whether it is correct: does it measure what it claims to measure?

Picture a bathroom scale that always shows two kilograms too much. Step on it ten times and it gives the same reading each time, so it is reliable. Yet every reading is wrong, so it is not valid. The misreading to avoid is to treat a consistent result as a true one: repeating a biased measure a thousand times only repeats the bias.

Three targets with five shots each: shots clustered on the center, labeled reliable and valid; shots clustered away from the center, labeled reliable, not valid; shots scattered over the target, labeled not reliable.Reliable and validReliable, not validNot reliable
A sketch: reliability is how close the shots are to each other; validity is how close they are to the center.

How reliability is checked

In the test-retest method the same questionnaire is given to the same people twice, a few weeks apart. If the trait it measures is stable, such as attitudes to work, the answers should be similar both times. Where people code what they see, researchers check inter-rater reliability: two observers code the same events separately and the share of agreement is counted. In the table, two observers coded five episodes in a classroom and agreed on four, so the agreement is 4 ÷ 5 = 80 percent. Low agreement means the categories are vague and the observers need clearer definitions.

Standard questions, closed answer options and trained interviewers all raise reliability, which is one reason large surveys use them. A questionnaire can also be checked for internal consistency: when several questions measure the same attitude, a person who strongly agrees with one should tend to agree with the others. If the answers do not hang together, some of the questions are measuring something else.

A table of five classroom episodes coded by two observers: helping and helping, teasing and teasing, teasing and ignoring, ignoring and ignoring, helping and helping. They agree on four of the five.EpisodeObserver AObserver BAgree?1HelpingHelpingyes2TeasingTeasingyes3TeasingIgnoringno4IgnoringIgnoringyes5HelpingHelpingyes
A made-up example: agreement on 4 of 5 episodes is 80 percent.

How validity is checked

The first test is face validity: on the face of it, does the indicator fit the concept? Income alone is a weak measure of social class, because class also involves education and the prestige of an occupation. Official crime statistics measure only crimes that are reported to the police and recorded. The rest, which criminologists call the dark figure of crime, is estimated by victimization surveys that ask people directly whether they were victims.

Two more kinds matter for studies of cause. Internal validity asks whether the conclusion holds inside the study: did the independent variable really cause the change, or did a third variable? External validity asks whether the result holds beyond it, for other people and places, which depends above all on a representative sample. A sociologist who wants to know how much a group reads would do well to combine indicators: library loans, books bought, e-books and a diary of one week's reading. Using several methods or indicators for one concept, called triangulation, is a common way to raise validity.

A table of three concepts, their indicators and the validity problem of each: crime, police records, only reported crimes; social class, income alone, ignores education; reading, library loans, misses e-books.ConceptIndicatorValidity problemCrimepolice recordsonly reported crimesSocial classincome aloneignores educationReadinglibrary loansmisses e-books
An indicator can be easy to count and still miss part of its concept.

How the two fit together

A measure can be reliable and not valid, like the biased scale. It cannot be valid without being reliable, because a tool whose results jump around at random cannot be hitting the truth. Reliability is therefore a condition for validity, not a proof of it. In practice the two can pull apart: closed questionnaires are highly reliable but may miss what answers mean, while in-depth interviews come closer to people's real experience but are harder to repeat. Official statistics show the tension well. A crime count compiled the same way every year is reliable and allows years to be compared, yet if people become more willing to report a crime, a rise in the count may reflect more reporting rather than more crime. A strong answer names both qualities and says which one a given method risks.

A two by two grid: reliable and valid is the goal; reliable and not valid is a biased scale; valid and not reliable is not possible; not reliable and not valid is random noise.ValidNot validReliablethe goala biased scaleNot reliablenot possiblerandom noise
Reliability is needed for validity, but does not guarantee it.
Free printable worksheet

Practice: reliability and validity

For 11th grade students of sociology. Answer each question from what is printed on the sheet, then check your work against the answers.

Free, no sign-up. The answer key prints on the second page. Regenerating a new worksheet on the same topic needs a free account.

Free printable worksheetlessoons.com
Practice: reliability and validityName:
  1. Write the label of the middle target.

    Three targets: shots on the center labeled reliable and valid; shots clustered off the center with a question mark; scattered shots labeled not reliable.Reliable and valid?Not reliable
  2. Reliability or validity problem? (a) Two interviewers asking the same questions get very different answers from the same people. (b) A study measures school success only by attendance. (c) A questionnaire on trust gives the same people very different scores a week apart.

  3. Fill in the last column, then work out the observers' agreement as a percentage.

    The table of five classroom episodes coded by observers A and B, with the agree column replaced by question marks.EpisodeObserver AObserver BAgree?1HelpingHelping?2TeasingTeasing?3TeasingIgnoring?4IgnoringIgnoring?5HelpingHelping?
  4. Fill in the validity problem of police records as a measure of crime, and name a way to estimate what they miss.

    The table of concepts and indicators with the validity problem of police records replaced by a question mark.ConceptIndicatorValidity problemCrimepolice records?Social classincome aloneignores educationReadinglibrary loansmisses e-books
  5. Complete the grid.

    The two by two grid of reliability and validity with the reliable-not-valid and valid-not-reliable cells replaced by question marks.ValidNot validReliablethe goal?Not reliable?random noise
  6. A study of one Tel Aviv school finds that a mentoring program raised grades. Two other changes were made in the school that year. Name one problem of internal validity and one of external validity.

  7. Design a way to measure how involved students are in class. Explain one step that makes it reliable and one that makes it valid.

More topics and worksheets at lessoons.com© 2026 Lessons

Common questions

Can research be reliable but not valid?

Yes. A measure can give the same result every time and still be wrong, like a scale that always adds two kilograms, or a questionnaire whose questions consistently measure something other than its topic.

Can research be valid but not reliable?

No. If a measure gives different results each time for the same thing, it cannot be measuring that thing accurately. Reliability is a condition for validity, though it does not guarantee it.

What is test-retest reliability?

A check of consistency over time: the same people answer the same questionnaire twice, a few weeks apart. Similar results both times show that the measure is reliable, as long as the trait itself is stable.

What is the difference between internal and external validity?

Internal validity asks whether a study's conclusion about cause holds within the study. External validity asks whether its results apply to other people, places and times beyond the sample.

Which is more important, reliability or validity?

Validity is the goal, since a study must measure what it claims to. But it cannot be reached without reliability, so researchers work on both, and choose methods knowing which one each method risks.

Still have a question? Ask Lessi

Any question about this subject. Lessi answers right here, step by step.

Try asking

Asking needs a free account; your question is kept while you sign up. Answers are written by AI and can contain mistakes. Lessi doesn’t give medical or legal advice.

Learn it with a teacher

Learn this topic live with Lessi

Lessi, the AI teacher, explains it out loud on a board, with short quizzes, at your pace. The first 60 minutes are free.