Round 1

Reviewer 1 Report

Examining the role of familiarity in the perception of depth.
Mischenko, Negishi, Gorbunova, Sawada

This paper investigates the effect of familiarity on perceived distance. It’s an old topic and the authors are right to stress how many mistakes are possible when studying it. They do a good job avoiding technical errors in their experiments, the cross-cultural control is clever, and the conclusions they draw from their data are justified.

One possible consequence of being careful with methods, though, is being overly careful, so that the topic that’s being studied ends up on the cutting-room floor, as it were, with the potential pitfalls. So, there are some loose ends here regarding the extent to which familiarity was given an adequate opportunity to influence observers’ perception of depth. The authors acknowledge this concern, briefly, in the Discussion, where, on p. 11, they note that, “It is possible that the effect of the familiarity cue on the depth perception may require a longer viewing duration.”

This could, of course, be tested. But, as explained on p.2, in the paragraph starting on line 67, the experiment was designed to avoid 'elaborate thinking' and instead tap into something we might thing of as raw, unmediated perception. But it’s not made clear what 'elaborate thinking' means. It's not distinguished from perception that just takes a long time. If 'thinking' means retrieving an accurate memory of a familiar object, then it’s exactly what the familiarity effect should depend on.

This bring us to another concern. The stimuli were presented for what should be enough time to identify a coin, especially one presented at fixation (Exp. 2). But the task does not require that the coin be identified as a coin, let alone a particular one. And no evidence was presented that the stimuli were recognized as coins either generically or as the particular coins they represented. Judging depth might involve analyzing object retinal size, but it doesn't necessitate analyzing object identity. Why should we assume that familiar coins were identified as familiar coins, rather than disks, in this task environment? It’s true that some of the coins differ in color and even shape. In some cases overall texture differences also exist. So, it might be convincing that at least some of them are uniquely specified on the basis of low-level visual properties, but we don’t know this. There are cities that are now virtually cashless societies, where familiarity with hard currency is a thing of the past. Coins are not felt and if they’re seen, they’re seen as coins printed on paper or a displayed on a screen. Most people who are likely to be enrolled in psychology experiments belong to representational cultures, where images are as commonly encountered as real objects. How does experience with images affect familiarity with the represented object? Does exposure to paintings, photographs and video affect our calibration of depth from accommodation and other cues? How well could the subjects of this experiment distinguish true from counterfeit coins that differ only in size? Clearly, these are issues that go beyond the simple question asked in this study but they are of direct concern to assumptions about familiarity that are implicitly made here. The authors make the point that “the failure to observe any familiarity effect can always be explained away simply by claiming that the observers did not have enough familiarity with the stimuli”, which is true but it cuts both ways: Assuming familiarity doesn’t make it so.

I think there’s a need to address these issues more fully in the Discussion, at least by acknowledging that there’s no direct evidence presented here about how familiar the subjects were with the stimuli, in particular the actual sizes of the stimuli, and about whether the coins were identified during task performance. I do think the data, within the constraints of the experiment, are not a problem. The problem is that the constraints might have determined the data. That might be unavoidable and, if acknowledged, is not necessarily a issue. It’s part of what makes this a difficult topic to study.

Other issues:

line 74: Not clear what ‘(p.18)’ might refer to in reference 11 (if it does refer to ref. 11). And the link for reference 11 doesn’t seem to work.

p. 2: In these 'could have' examples (line 75 parag), it would be nice to have specifics of what could go wrong. Just because some confounding could occur doesn't mean it's sufficient or even plausible as an account of experimental results. Listing confounding variables, as done here, is fine, but giving an example or two of how it would work out in practice would give some substance to this discussion.

Line 322: define N1/2

Fig. 2: Different numbers of points appear in the two figures, A and B. Yet one figure is a replot of one of the two sets of values presented in the other. This is confusing—until it’s understood that what appear to be small s.d. values in A are in fact missing data, those of Gogle 1969b, as is clear from the Supplement. (Also, Gogle 1969a and 1969b should be distinguished in the figure legend.)

Sect. 2.3:
- It’s not clear what conclusion is to be drawn from this section. In a way it summarizes a portion of the literature, so it’s relevant, but it’s not integrated with anything else in the paper. If it is meant as a condensation of support for a familiarity effect, it should be identified as such.

Lines 202-4: “The probability correct was higher (i) when it was based on the depth responses (Equation 3) than (ii) when it was based on the difference between the distance responses (Equation 4).”
- But Fig. 4 shows the opposite. Probably a labeling reversal.

Author Response

Thank you very much for reviewing our manuscript and for the constructive feedback. We are happy to see your interest to our study. We revised our manuscript with taking all your suggestions into account as detailed point by point below. All the revisions are in red with balloon comments in the revised manuscript submitted as a supplement file.

R1a: One possible consequence of being careful with methods, though, is being overly careful, so that the topic that’s being studied ends up on the cutting-room floor, as it were, with the potential pitfalls. So, there are some loose ends here regarding the extent to which familiarity was given an adequate opportunity to influence observers’ perception of depth. The authors acknowledge this concern, briefly, in the Discussion, where, on p. 11, they note that, “It is possible that the effect of the familiarity cue on the depth perception may require a longer viewing duration.”

We agree with the reviewer. We expanded a discussion about limitations of the study in 4. Discussion section (p.12, l.396, 401, 408).

R1b: This could, of course, be tested. But, as explained on p.2, in the paragraph starting on line 67, the experiment was designed to avoid 'elaborate thinking' and instead tap into something we might thing of as raw, unmediated perception. But it’s not made clear what 'elaborate thinking' means. It's not distinguished from perception that just takes a long time. If 'thinking' means retrieving an accurate memory of a familiar object, then it’s exactly what the familiarity effect should depend on.

For example, the judgment may be based on solving Equation 1 consciously rather than on the distance and depth perceived. We added this sentence and references discussing the effect of the thinking or of other cognitive bias on the judgement (p.2, l.72). We also added text discussing the retrieval of memorized size (p.12, l.401).

R1c: This bring us to another concern. The stimuli were presented for what should be enough time to identify a coin, especially one presented at fixation (Exp. 2). But the task does not require that the coin be identified as a coin, let alone a particular one. And no evidence was presented that the stimuli were recognized as coins either generically or as the particular coins they represented. Judging depth might involve analyzing object retinal size, but it doesn't necessitate analyzing object identity. Why should we assume that familiar coins were identified as familiar coins, rather than disks, in this task environment? It’s true that some of the coins differ in color and even shape. In some cases overall texture differences also exist. So, it might be convincing that at least some of them are uniquely specified on the basis of low-level visual properties, but we don’t know this. There are cities that are now virtually cashless societies, where familiarity with hard currency is a thing of the past. Coins are not felt and if they’re seen, they’re seen as coins printed on paper or a displayed on a screen. Most people who are likely to be enrolled in psychology experiments belong to representational cultures, where images are as commonly encountered as real objects. How does experience with images affect familiarity with the represented object? Does exposure to paintings, photographs and video affect our calibration of depth from accommodation and other cues? How well could the subjects of this experiment distinguish true from counterfeit coins that differ only in size? Clearly, these are issues that go beyond the simple question asked in this study but they are of direct concern to assumptions about familiarity that are implicitly made here. The authors make the point that “the failure to observe any familiarity effect can always be explained away simply by claiming that the observers did not have enough familiarity with the stimuli”, which is true but it cuts both ways: Assuming familiarity doesn’t make it so.

We added a new paragraph addressing these issues in 3.1.3. Stimuli (p.8, l.269) and sentences in 4. Discussion (p.12, l.396, 408). We also added a list of familiar objects used in the past studies to supplemental material (Mischenko et al - Supplement.xlsx).

Coins are still commonly used in Moscow, Russia and in Kanazawa, Japan, especially by undergraduate students. We could use only cash in a canteen of in the Department of Psychology at the Higher School of Economics until recently in Russia. Also, it is still difficult for undergraduate students in Japan to have credit cards. Note that the past studies reported the familiarity effect using objects that were not so common as coins.

R1d: I think there’s a need to address these issues more fully in the Discussion, at least by acknowledging that there’s no direct evidence presented here about how familiar the subjects were with the stimuli, in particular the actual sizes of the stimuli, and about whether the coins were identified during task performance. I do think the data, within the constraints of the experiment, are not a problem. The problem is that the constraints might have determined the data. That might be unavoidable and, if acknowledged, is not necessarily a issue. It’s part of what makes this a difficult topic to study.

We added some text discussing limitations of this study (p.12, l.396, 401, 408). Note that the observers were informed before the experiments that the images of the coins would be shown as visual stimuli. A sentence explaining this instruction was added (p.8, l.286).

R1e: line 74: Not clear what ‘(p.18)’ might refer to in reference 11 (if it does refer to ref. 11). And the link for reference 11 doesn’t seem to work.

The text is revised so that p.18 of [16] (it was [11]) is referenced in this sentence (p.2, l.75). We also updated the link of [16] in References section (p.13, l.457).

R1f: p. 2: In these 'could have' examples (line 75 parag), it would be nice to have specifics of what could go wrong. Just because some confounding could occur doesn't mean it's sufficient or even plausible as an account of experimental results. Listing confounding variables, as done here, is fine, but giving an example or two of how it would work out in practice would give some substance to this discussion.

We added an explanation of one example in a footnote (p.2, l.81; p.2, Footnote 2).

R1g: Line 322: define N_1/2

Thank you for pointing it out. We added a sentence defining N_1/2 (p.10, l.332).

R1h: Fig. 2: Different numbers of points appear in the two figures, A and B. Yet one figure is a replot of one of the two sets of values presented in the other. This is confusing—until it’s understood that what appear to be small s.d. values in A are in fact missing data, those of Gogle 1969b, as is clear from the Supplement. (Also, Gogle 1969a and 1969b should be distinguished in the figure legend.)

We added sentences in captions of Figures 2 (p.4, l.145, 147) and 3 (p.5, l.177) to clarify these points. We also added reference numbers in a legend of Figure 2 (p.4).

R1i: Sect. 2.3: It’s not clear what conclusion is to be drawn from this section. In a way it summarizes a portion of the literature, so it’s relevant, but it’s not integrated with anything else in the paper. If it is meant as a condensation of support for a familiarity effect, it should be identified as such.

At the end of Section 2.3, we added a sentence connecting this section with the psychophysical experiments (p.6, l.208).

R1j: Lines 202-4: “The probability correct was higher (i) when it was based on the depth responses (Equation 3) than (ii) when it was based on the difference between the distance responses (Equation 4).” 
- But Fig. 4 shows the opposite. Probably a labeling reversal.

Thank you for pointing it out. This error in Figure 4 is corrected.

The other reviewer (Reviewer-2)

R2a: The meta analysis of the exisitng literature is well-performed and leads to suggest that familiarity may influence depth perception. Such an effect is, however, very likely to depend on the type of "familiar" material selected for the tests. We definitely do not feel "familiar" with the faces of friends, aquaintances, and close family in the same way as we may feel "familiar" with the coins of our currency, with which we share no memories or experience. My point here is: the effects previously found in other experiments may well be explained by a specific emotional charge associated with the familiar material used for testing. There is evidence in the literature that strong emotions may, indeed, alter our perceptions and this may include the perception of the relative depth of familiar objects. When we see the face of a family member, friend, or animal, for example, we feel something special, and our perception is inevitably associated with special memories and events relaed to that person or animal (just to gvie one example - we may also have strong emotions associate with other familiar objects). When we see a familiar piece of money we use every day, such as a Eurocoin or a Dollar, it is unlikely that there is any particular or significant emotion associated with that perception.

In this study, being familiar means that an object commonly appears in our everyday life. There are some studies testing any effect of an emotional attachment on an object but we do not consider the emotional effect in this study. We added a new footnote with references to the studies of the emotional effect to discuss this difference (p.2, footnote 1).

R2b: In short: The revised manscript needs to provide a clear and critical discussion of the potential importance of the type of familiar stimulus material selected for testing in the subjective depth experiment - in the introduction and the results/discussion/conclusions sections.

We added a new paragraph addressing this issue in 3.1.3. Stimuli (p.8, l.269) and made a list of familiar objects used in the past studies to supplemental material (Mischenko et al - Supplement.xlsx).

R2c: English grammar and style need to be carefully checked, line by line, some sentences are not very clear/clumsily formulated.

The text of the manuscript was checked by a native English speaker.

Additional revision

The number of Russian and Japanese participants was registered to be N_1/2 = 40 (https://osf.io/n5czq), but the pre-registration also reported that the estimated probability for observing a familiarity effect with N_1/2 = 50. This inconsistency of N_1/2 is our mistake. We stated this in an additional footnote (p.10, footnote 3).

Reviewer 2 Report

This original research article discusses data from a multiple choice experiment where subjects had to assess the relative depth of pieces of money (coins) to determine whether familiarity with the coins would influence their depth judgments, as is suggested by findings from other experiments with other types of stimuli, published previously. The study described here produced no effect of familiarity on depth judgments. This is an interesting and novel negative finding that definitely warrants publication. However, the authors are urged to perform some revisions along the following lines:

The meta analysis of the exisitng literature is well-performed and leads to suggest that familiarity may influence depth perception. Such an effect is, however, very likely to depend on the type of "familiar" material selected for the tests. We definitely do not feel "familiar" with the faces of friends, aquaintances, and close family in the same way as we may feel "familiar" with the coins of our currency, with which we share no memories or experience. My point here is: the effects previously found in other experiments may well be explained by a specific emotional charge associated with the familiar material used for testing. There is evidence in the literature that strong emotions may, indeed, alter our perceptions and this may include the perception of the relative depth of familiar objects. When we see the face of a family member, friend, or animal, for example, we feel something special, and our perception is inevitably associated with special memories and events relaed to that person or animal (just to gvie one example - we may also have strong emotions associate with other familiar objects). When we see a familiar piece of money we use every day, such as a Eurocoin or a Dollar, it is unlikely that there is any particular or significant emotion associated with that perception.

In short: The revised manscript needs to provide a clear and critical discussion of the potential importance of the type of familiar stimulus material selected for testing in the subjective depth experiment - in the introduction and the results/discussion/conclusions sections.

English grammar and style need to be carefully checked, line by line, some sentences are not very clear/clumsily formulated.

Author Response

Thank you very much for reviewing our manuscript and for the positive feedback. We are happy to see your interest to our study. We revised our manuscript with taking all your suggestions into account as detailed point by point below. All the revisions are in red with balloon comments in the revised manuscript submitted as a supplement file.

R2a: The meta analysis of the exisitng literature is well-performed and leads to suggest that familiarity may influence depth perception. Such an effect is, however, very likely to depend on the type of "familiar" material selected for the tests. We definitely do not feel "familiar" with the faces of friends, aquaintances, and close family in the same way as we may feel "familiar" with the coins of our currency, with which we share no memories or experience. My point here is: the effects previously found in other experiments may well be explained by a specific emotional charge associated with the familiar material used for testing. There is evidence in the literature that strong emotions may, indeed, alter our perceptions and this may include the perception of the relative depth of familiar objects. When we see the face of a family member, friend, or animal, for example, we feel something special, and our perception is inevitably associated with special memories and events relaed to that person or animal (just to gvie one example - we may also have strong emotions associate with other familiar objects). When we see a familiar piece of money we use every day, such as a Eurocoin or a Dollar, it is unlikely that there is any particular or significant emotion associated with that perception.

In this study, being familiar means that an object commonly appears in our everyday life. There are some studies testing any effect of an emotional attachment on an object but we do not consider the emotional effect in this study. We added a new footnote with references to the studies of the emotional effect to discuss this difference (p.2, footnote 1).

R2b: In short: The revised manscript needs to provide a clear and critical discussion of the potential importance of the type of familiar stimulus material selected for testing in the subjective depth experiment - in the introduction and the results/discussion/conclusions sections.

We added a new paragraph addressing this issue in 3.1.3. Stimuli (p.8, l.269) and made a list of familiar objects used in the past studies to supplemental material (Mischenko et al - Supplement.xlsx).

R2c: English grammar and style need to be carefully checked, line by line, some sentences are not very clear/clumsily formulated.

The text of the manuscript was checked by a native English speaker.

The other reviewer (Reviewer-1)

R1a: One possible consequence of being careful with methods, though, is being overly careful, so that the topic that’s being studied ends up on the cutting-room floor, as it were, with the potential pitfalls. So, there are some loose ends here regarding the extent to which familiarity was given an adequate opportunity to influence observers’ perception of depth. The authors acknowledge this concern, briefly, in the Discussion, where, on p. 11, they note that, “It is possible that the effect of the familiarity cue on the depth perception may require a longer viewing duration.”

We agree with the reviewer. We expanded a discussion about limitations of the study in 4. Discussion section (p.12, l.396, 401, 408).

R1b: This could, of course, be tested. But, as explained on p.2, in the paragraph starting on line 67, the experiment was designed to avoid 'elaborate thinking' and instead tap into something we might thing of as raw, unmediated perception. But it’s not made clear what 'elaborate thinking' means. It's not distinguished from perception that just takes a long time. If 'thinking' means retrieving an accurate memory of a familiar object, then it’s exactly what the familiarity effect should depend on.

For example, the judgment may be based on solving Equation 1 consciously rather than on the distance and depth perceived. We added this sentence and references discussing the effect of the thinking or of other cognitive bias on the judgement (p.2, l.72). We also added text discussing the retrieval of memorized size (p.12, l.401).

R1c: This bring us to another concern. The stimuli were presented for what should be enough time to identify a coin, especially one presented at fixation (Exp. 2). But the task does not require that the coin be identified as a coin, let alone a particular one. And no evidence was presented that the stimuli were recognized as coins either generically or as the particular coins they represented. Judging depth might involve analyzing object retinal size, but it doesn't necessitate analyzing object identity. Why should we assume that familiar coins were identified as familiar coins, rather than disks, in this task environment? It’s true that some of the coins differ in color and even shape. In some cases overall texture differences also exist. So, it might be convincing that at least some of them are uniquely specified on the basis of low-level visual properties, but we don’t know this. There are cities that are now virtually cashless societies, where familiarity with hard currency is a thing of the past. Coins are not felt and if they’re seen, they’re seen as coins printed on paper or a displayed on a screen. Most people who are likely to be enrolled in psychology experiments belong to representational cultures, where images are as commonly encountered as real objects. How does experience with images affect familiarity with the represented object? Does exposure to paintings, photographs and video affect our calibration of depth from accommodation and other cues? How well could the subjects of this experiment distinguish true from counterfeit coins that differ only in size? Clearly, these are issues that go beyond the simple question asked in this study but they are of direct concern to assumptions about familiarity that are implicitly made here. The authors make the point that “the failure to observe any familiarity effect can always be explained away simply by claiming that the observers did not have enough familiarity with the stimuli”, which is true but it cuts both ways: Assuming familiarity doesn’t make it so.

We added a new paragraph addressing these issues in 3.1.3. Stimuli (p.8, l.269) and sentences in 4. Discussion (p.12, l.396, 408). We also added a list of familiar objects used in the past studies to supplemental material (Mischenko et al - Supplement.xlsx).

Coins are still commonly used in Moscow, Russia and in Kanazawa, Japan, especially by undergraduate students. We could use only cash in a canteen of in the Department of Psychology at the Higher School of Economics until recently in Russia. Also, it is still difficult for undergraduate students in Japan to have credit cards. Note that the past studies reported the familiarity effect using objects that were not so common as coins.

R1d: I think there’s a need to address these issues more fully in the Discussion, at least by acknowledging that there’s no direct evidence presented here about how familiar the subjects were with the stimuli, in particular the actual sizes of the stimuli, and about whether the coins were identified during task performance. I do think the data, within the constraints of the experiment, are not a problem. The problem is that the constraints might have determined the data. That might be unavoidable and, if acknowledged, is not necessarily a issue. It’s part of what makes this a difficult topic to study.

We added some text discussing limitations of this study (p.12, l.396, 401, 408). Note that the observers were informed before the experiments that the images of the coins would be shown as visual stimuli. A sentence explaining this instruction was added (p.8, l.286).

R1e: line 74: Not clear what ‘(p.18)’ might refer to in reference 11 (if it does refer to ref. 11). And the link for reference 11 doesn’t seem to work.

The text is revised so that p.18 of [16] (it was [11]) is referenced in this sentence (p.2, l.75). We also updated the link of [16] in References section (p.13, l.457).

R1f: p. 2: In these 'could have' examples (line 75 parag), it would be nice to have specifics of what could go wrong. Just because some confounding could occur doesn't mean it's sufficient or even plausible as an account of experimental results. Listing confounding variables, as done here, is fine, but giving an example or two of how it would work out in practice would give some substance to this discussion.

We added an explanation of one example in a footnote (p.2, l.81; p.2, Footnote 2).

R1g: Line 322: define N_1/2

Thank you for pointing it out. We added a sentence defining N_1/2 (p.10, l.332).

R1h: Fig. 2: Different numbers of points appear in the two figures, A and B. Yet one figure is a replot of one of the two sets of values presented in the other. This is confusing—until it’s understood that what appear to be small s.d. values in A are in fact missing data, those of Gogle 1969b, as is clear from the Supplement. (Also, Gogle 1969a and 1969b should be distinguished in the figure legend.)

We added sentences in captions of Figures 2 (p.4, l.145, 147) and 3 (p.5, l.177) to clarify these points. We also added reference numbers in a legend of Figure 2 (p.4).

R1i: Sect. 2.3: It’s not clear what conclusion is to be drawn from this section. In a way it summarizes a portion of the literature, so it’s relevant, but it’s not integrated with anything else in the paper. If it is meant as a condensation of support for a familiarity effect, it should be identified as such.

At the end of Section 2.3, we added a sentence connecting this section with the psychophysical experiments (p.6, l.208).

R1j: Lines 202-4: “The probability correct was higher (i) when it was based on the depth responses (Equation 3) than (ii) when it was based on the difference between the distance responses (Equation 4).” 
- But Fig. 4 shows the opposite. Probably a labeling reversal.

Thank you for pointing it out. This error in Figure 4 is corrected.

Additional revision

The number of Russian and Japanese participants was registered to be N_1/2 = 40 (https://osf.io/n5czq), but the pre-registration also reported that the estimated probability for observing a familiarity effect with N_1/2 = 50. This inconsistency of N_1/2 is our mistake. We stated this in an additional footnote (p.10, footnote 3).