How good research earns confidence across quantitative, qualitative, and mixed methods designs

1. Intuitive Introduction

Researchers often talk about “good evidence” as if its quality were obvious. But evidence is never strong simply because it exists. It becomes strong when the study gives readers good reasons to trust the connection between the evidence, the interpretation, and the claim. In quantitative work, this discussion often appears through the language of validity. In qualitative work, it often appears through terms such as credibility and trustworthiness. In mixed methods research, it also includes the quality of integration and the defensibility of combined inference. Creswell and Creswell’s sixth edition presents design quality as a matter of fit among question, procedures, evidence, and interpretation, while Lincoln and Guba’s classic Naturalistic Inquiry helped establish trustworthiness as a central qualitative concern.

This blog post therefore asks a practical question: when are we justified in trusting what a study concludes? That question looks slightly different across research traditions, but the underlying issue is shared. A strong study does not merely collect data. It earns confidence. Shadish, Cook, and Campbell’s design framework makes this point for quantitative inference, while Morse, Barrett, Mayan, Olson, and Spiers argue that rigor in qualitative research must also be actively established rather than assumed.

2. Why This Post Matters

This post comes at a crucial point in the blog post mini-series. Earlier posts focused on building a study: formulating questions, choosing methodology, translating concepts into evidence, and planning sampling or case selection. But even a well-structured study still faces a deeper test: can its findings be trusted? That is where many otherwise promising projects begin to weaken. The question is no longer only whether the study is well organized. It is whether the design, evidence, and interpretation are strong enough to justify confidence.

This is why the post is not just about terminology. It is about methodological maturity. Readers often inherit the mistaken idea that quantitative work is judged by validity, qualitative work by trustworthiness, and mixed methods by integration, as if these were separate worlds. In reality, all three are trying to solve a related problem: how to justify the quality of inference. The vocabulary differs because the inferential tasks differ, but the burden of justification remains. Creswell and Creswell present all three design families as distinct but comparable forms of inquiry, and Lincoln and Guba’s work reminds us that qualitative alternatives to conventional validity language were created to address that same burden in a different epistemological register.

3. The Main Logic of This Post

The dominant logic here is M > D > C, where C stands for the final claim or conclusion. Methodology comes first because the design determines what kind of quality criteria are relevant. Data come second because weak or thin evidence limits what can be trusted regardless of the design label. Claim comes third because the final interpretation must remain proportionate to both the design and the evidence.

This ranking is useful because readers sometimes start at the wrong end. They ask whether a conclusion sounds persuasive before asking whether the design was capable of supporting it. But a study earns confidence in the opposite direction. First the design must create the right conditions for quality. Then the evidence must be good enough for the intended inference. Only then does the final claim deserve trust.

4. Validity in Quantitative Research

In quantitative research, validity is often the central quality language because quantitative studies typically aim to measure, estimate, compare, predict, or infer with structured evidence. The concept is broad, but at its core validity asks whether the study is getting the right answer to the question it actually set out to address. Shadish, Cook, and Campbell emphasize that strong quantitative inference depends on design, not merely on statistical procedure. Their work remains foundational because it treats internal validity, external validity, construct validity, and conclusion validity as interconnected design concerns rather than isolated technical checks.

This matters because weak quantitative studies often fail long before the analysis. If the constructs are poorly translated into variables, construct validity weakens. If comparison groups are not meaningfully comparable, internal validity weakens. If the sample is too narrow for the generalization being made, external validity weakens. If the statistical interpretation outruns what the design can support, the strength of the conclusion weakens. In other words, validity is not one box to be ticked at the end. It is a layered judgment about whether the study design earns confidence.

5. Credibility and Trustworthiness in Qualitative Research

Qualitative research often uses different quality language because its goals are different. A qualitative study is usually not trying to estimate a population parameter or test a variable-based hypothesis in the same way as a quantitative one. It is often trying to understand meaning, process, context, practice, or lived experience. For this reason, Lincoln and Guba proposed the language of trustworthiness, including credibility, transferability, dependability, and confirmability, as an alternative to simply importing quantitative criteria unchanged into qualitative inquiry. Their 1985 book remains one of the key sources for that framework.

Credibility, in this context, asks whether the interpretation is well grounded in the evidence. Trustworthiness asks more broadly whether the study gives readers good reasons to believe that the findings were produced carefully, reflexively, and transparently. Morse and colleagues argued that rigor in qualitative work is not achieved by ritual declarations, but by verification strategies built into the design and conduct of the study itself. That point is especially important because some qualitative studies borrow the language of trustworthiness while still relying on thin evidence, vague procedures, or inflated interpretation. The language alone does not make a study trustworthy. The design has to earn it.

6. Quality of Inference in Mixed Methods Research

Mixed methods research introduces a further layer of quality because it must justify not only the credibility of each strand, but also the defensibility of their integration. A mixed methods study can fail even when the quantitative and qualitative components are individually competent if the connection between them is weak or merely decorative. Fetters, Curry, and Creswell argue that integration is the core design issue in mixed methods research, and that the quality of inference depends on how effectively the strands are brought into relation at the levels of design, methods, interpretation, or reporting.

This means that mixed methods quality is not just the sum of two separate standards. It is also a question of whether the combined inference is stronger because of the integration. If the quantitative strand estimates patterns and the qualitative strand explores meanings, the study should show what becomes newly understandable when those two forms of evidence are linked. If that added value never appears, then the study may still be multimethod, but it is not especially strong as mixed methods inquiry.

7. One Problem, Several Quality Languages

At first glance, validity, credibility, trustworthiness, and quality of inference can seem like competing methodological vocabularies. But they are better understood as different responses to the same underlying question: why should the reader trust this study? The answer cannot be identical across all designs because the inferential tasks are not identical. A randomized trial, an ethnographic study, and an explanatory sequential mixed methods design do not earn confidence in exactly the same way. Yet they all face the same burden of justification.

This is an important educational point. Students often become confused because they think the existence of different vocabularies means that standards disappear across designs. In fact, the opposite is true. Different designs require different standards precisely because quality must be judged in relation to the kind of knowledge the study seeks to produce. The common thread is not one universal checklist, but one universal obligation: to justify confidence in the findings and the inference.

8. Common Threats to Quality Across Designs

Despite the differences in language, many threats to quality are recognizable across research traditions. One threat is misalignment: the study asks one kind of question but uses evidence more suited to another. A second threat is thin evidence: the data exist, but they are too weak, too narrow, or too indirect for the claim. A third threat is interpretive inflation: the conclusion says more than the evidence and design can support. A fourth threat is token procedure: quality language is used ceremonially rather than built into the study architecture.

These threats matter because they show that quality is not mainly about labels. A study can say “valid,” “credible,” or “trustworthy” and still fail to earn those descriptions. Real quality appears when the design anticipates threats and responds to them honestly and transparently.

9. How to Build Quality Into the Design Early

The best way to improve quality is not to add defensive language at the end of the paper. It is to build quality into the study from the start. In quantitative work, that means thinking early about validity threats, comparability, measurement, and inferential scope. In qualitative work, it means designing for information richness, contextual depth, reflexive transparency, and credible interpretation rather than trying to rescue weak material afterward. In mixed methods work, it means planning how integration will occur rather than hoping it will emerge naturally.

This chapter therefore reinforces a broader lesson from the book: design quality is upstream. By the time a researcher is writing the limitations section, many quality strengths or weaknesses are already built in.

10. What Can Still Be Repaired Later

Some quality problems can still be improved after data collection, but many can only be repaired partially. A quantitative study with narrow data may still become stronger if the claims are reduced and the limits of inference made explicit. A qualitative study with thinner-than-planned material may still yield a narrower, more descriptive contribution if the interpretation is disciplined accordingly. A mixed methods study may still improve through better integration at the analysis stage, even if the original design was not perfect.

But not everything can be saved. A badly measured construct cannot be fully repaired by elegant statistics. A generic, low-trust interview cannot be turned into deep qualitative evidence by coding alone. Two disconnected strands do not become a strong mixed methods design simply because they are discussed together. This is why the chapter places so much emphasis on prevention rather than rescue.

11. Methodological Maturity and Earned Confidence

Methodological maturity begins when researchers stop treating quality language as decoration and start treating it as a burden of proof. A mature study does not merely announce validity, credibility, or trustworthiness. It shows how those were earned. It also accepts that not all studies can support the same kind of confidence. Some studies support local, bounded, exploratory insight. Others support broader or more robust inference. Strength lies not in claiming the highest possible level of confidence, but in honestly matching the claim to what the design and evidence deserve.

That is the central lesson of this post. Good research is not simply about producing findings. It is about producing findings that readers have reason to trust.

Mini-Checklist

Before finalizing a study, ask:

• What quality language is appropriate for this design: validity, credibility, trustworthiness, integration quality, or some combination?
• Have I actually earned that language through the design and evidence, or only used it rhetorically?
• What are the main threats to confidence in this study?
• Does my final claim stay within the level of trust the design can justify?
• If a skeptical reader asked, “Why should I trust this study?”, could I answer clearly and specifically?

A strong study is not the one that uses the most impressive quality vocabulary. It is the one that truly earns confidence.

References

Creswell, J. W., & Creswell, J. D. (2023). Research design: Qualitative, quantitative, and mixed methods approaches (6th ed.). SAGE.

Fetters, M. D., Curry, L. A., & Creswell, J. W. (2013). Achieving integration in mixed methods designs—Principles and practices. Health Services Research, 48(6 Pt 2), 2134–2156. https://doi.org/10.1111/1475-6773.12117

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. SAGE.

Maxwell, J. A. (2013). Qualitative research design: An interactive approach (3rd ed.). SAGE.

Morse, J. M., Barrett, M., Mayan, M., Olson, K., & Spiers, J. (2002). Verification strategies for establishing reliability and validity in qualitative research. International Journal of Qualitative Methods, 1(2), 13–22. https://doi.org/10.1177/160940690200100202

Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasi-experimental designs for generalized causal inference. Houghton Mifflin.