Mistyping between INFJ and INFP is one of the most-documented patterns in MBTI literature. Two specific patterns account for most of it.
**Pattern 1: INFP self-typing as INFJ.** This is the most common direction. An INFP who has internalized the popular characterization of INFJ as "rare, deeply empathic, mystical insight type" reads themselves into the description because the introspective sensitivity matches. The test: when you make a hard decision, do you check it against your personal values (Fi → INFP) or against the impact on the group / others (Fe → INFJ)? When you give difficult feedback, do you weight "is this true to what I believe" (Fi → INFP) or "will this preserve / improve the relationship" (Fe → INFJ)? Under stress, do you become more harsh and forcibly-organized (Te → INFP inferior) or more disconnected from your body and overindulgent (Se → INFJ inferior)?
**Pattern 2: INFJ self-typing as INFP.** Less common but real. An INFJ with strong personal values may read themselves as Fi-led because they identify with the values-language. The test: do other people's emotional states show up in your awareness without you trying to track them (Fe → INFJ), or do you have to consciously direct attention to read others (less natural Fe → INFP)? Do you tend to know what a group needs before they articulate it (Ni + Fe → INFJ), or do you tend to have a strong personal stance that you then need to translate for others (Fi + Ne → INFP)?
**A useful self-test:** read the type profile for INFJ and INFP back-to-back. Then read INTJ and INFP back-to-back, and INFJ and ENFP back-to-back. The type whose differential signal feels strongest — i.e., where one description feels distinctly more accurate than its neighbor — is more likely to be your actual type. If both INFJ and INFP feel equally accurate to you, you may be reading Forer-effect prose rather than type-specific signal (per Forer 1949, DOI 10.1037/h0059240, generic descriptions feel personally accurate to most people regardless of any underlying type).
**A useful structural caveat:** if you've taken MBTI multiple times and switched between INFJ and INFP, you're almost certainly near the midpoint on at least one dimension (most likely J/P). Per Pittenger 2005 (DOI 10.1037/1065-9293.57.3.210), MBTI's per-dimension test-retest reliability is around 0.5–0.6, which means scores near the cutoff flip easily on retest. The per-dimension stability is lower than Big Five (~0.7–0.9). If you keep switching letters, that letter is your weak signal, not a sign your personality is unstable.
For the longer treatment of MBTI's measurement limits and what "weak signal" actually means, see /blog/mbti-common-misconceptions-and-data — it covers the four primary citations (Pittenger, McCrae & Costa, Barrick & Mount, Forer) in depth.