Talk:Retrieval-induced forgetting

Comment

This is an impressive article- kudos to Jethrobot for doing so much work on it! The only criticism I would make is that the language is at a very technical level throughout: fine for an academic audience, but needs simplifying for Wikipedia's general audience. Always consider whether a long word or a technical term is absolutely necessary. Cheers, MartinPoulter (talk) 14:53, 23 September 2012 (UTC)

• Yeah...I'd like to think it is at least marginally better than how it is discussed in academic journals, but I agree that it is still very technical. I can work on making it more accessible, but sometimes even longer words (but not necessarily technical term) are necessary to be precise. I will probably tag the page for now given your opinion, and will work on fixing this over the next few weeks. I, Jethrobot ^{drop me a line} (note: not a bot!) 15:46, 23 September 2012 (UTC)

I think, however, the language in the lead is accessible to most readers. I, Jethrobot ^{drop me a line} (note: not a bot!) 15:51, 23 September 2012 (UTC)

"it is at least marginally better than how it is discussed in academic journals" - absolutely! "sometimes even longer words... are necessary to be precise"- agree as well. We should never sacrifice accuracy, and I leave up to you to make the judgement calls.

Re the lead, is the adjective "empirically" necessary in that context? "the plausibility of such an effect was previously alluded to" there must be a plainer way to say this. "inhibitory mechanism" - I know what this is but can we have a simpler phrase for newcomers. MartinPoulter (talk) 16:08, 23 September 2012 (UTC)

Nice work! Big improvement --MartinPoulter (talk) 09:32, 1 October 2012 (UTC)

Martin, do you have any strong opinions on whether, in the paradigm section, whether item types should be designated as their abbreviations (Rp+/Rp-/NRp) or their long-form, descriptive names (practiced items / unpracticed, related items / unpracticed, unrelated items)? There are obvious disadvantages for both of these. The first has the problem of coming off as jargon, which is what I am trying to fix with my edits recently. However, the second has a comparable problem: That such terms are long and convoluted, even if they are precise. I've struggled with this in academic papers I've written, but the conflict is felt much more here on Wikipedia. I, Jethrobot ^{drop me a line} (note: not a bot!) 04:49, 14 October 2012 (UTC)

Thanks for seeking my opinion. I strongly favour the long-form descriptive names. If Rp+ and Rp- occurred a lot more frequently, then I would pause. However, the number of times they occur (especially in the paradigm section) is small enough that I don't think the extra length would be an annoyance. Cheers, MartinPoulter (talk) 11:14, 14 October 2012 (UTC)

GA Review

This review is transcluded from Talk:Retrieval-induced forgetting/GA1. The edit link for this section can be used to add comments to the review.

Reviewer: Churn and change (talk · contribs) 04:11, 5 November 2012 (UTC)

GA review – see WP:WIAGA for criteria

1. Is it reasonably well written?

   A. Prose quality:

   General: technical terms are often handled via wikilinks. It is better to provide a brief description inline, and then wikilink. I will detail this for each section separately.

   Lead:

   • "explicit retrieval": add a short description of what it is, here (conscious remembering is good enough for a lay person). Done
   • Avoid starting sentences with "there are" or "it is." Psychologists have multiple theories on why RIF occcurs, and how it relates to the larger issue of cognition."  Done
   • "mechanism or process": just one or the other should be good.  Done

   Prior and related research:

   • The quote had better be paraphrased, since it is heavy on jargon: "prime," "semantically associated" and so on.  Done
   • part-set cueing needs a short description, not just a wikilink.  Done
   • I cut down on this sentence, but still have trouble with the "shown" part: shown what? A map?  Done
   • changes in one's environment looks better, though you can still wikilink to context effects.
     Resolved

   The problem with using "environment" is that I think it implies a physical change, where as "context" is more broad. In the cited study by Sahakyan, their manipulation of context was to instruct one group to think of early memories related to home, whereas another group was given a "forget" instruction for the list. There wasn't really a change in the environment. I, Jethrobot ^{drop me a line} (note: not a bot!) 08:10, 6 November 2012 (UTC)

   • Need a short explanation of perseveration.  Done

   Retrieval-practice paradigm:

   • Too many short paragraphs. The one longer para, retrieval-practice phase should probably be summarized leaving out detail such as how many repetitions occur and so on. I think the second para with the three bulleted items and the figure is the core. A lot of the first para and the examples look unnecessary and have a bit of a tutorial quality to it.  Done
   • The "test phase" need not have subsections; each type can just be a separate paragraph.  Done

   General findings:

   • What exactly does "baseline" mean here? I thought this just compares two different sets (haven't read the refs yet).  Done

   The "baseline" refers to the recall of unpracticed-unrelated (NRp) items. Basically, the question is, does retrieving some items from a given set of categories negatively affect related items any more than unrelated items? So, the unrelated items serve as a baseline, as the idea is that related items will more impaired as a result of retrieval practice. I, Jethrobot ^{drop me a line} (note: not a bot!)

   • The corollary, that "practiced words are remembered better at test time" seems obvious.
     Resolved

   I agree that it is obvious-- do you think it should be removed? It's not especially relevant (IMO) to the discussion of RIF, but it is a robust effect that arises regularly in these studies. I, Jethrobot ^{drop me a line} (note: not a bot!)

   I think it should be removed as something expected to occur regularly. Churn and change (talk) 17:47, 11 November 2012 (UTC)

   • I thought the "testing effect" was the better memory for something tested, because the testing invokes retrieval?
     Resolved

   Correct. A subset of items are tested during retrieval practice and, because all items are retested during the final phase, they get the benefit of prior testing. I, Jethrobot ^{drop me a line} (note: not a bot!)

   This too may show up in RIF tests, just as many other memory-related effects would. I don't think it is relevant to the subject though. Churn and change (talk) 17:47, 11 November 2012 (UTC)

   For now, I agree it is not the most relevant and so it has been removed. However, with more improvements to the theory section, there are some who compare retrieval practice with success with individual RIF performance as a way of testing interference accounts. So, I may include this information at a later time under the theories section. I, Jethrobot ^{drop me a line} (note: not a bot!) 22:12, 11 November 2012 (UTC)

   • Generality of RIF: "homographs", "lexical properties" and "propositional information" all need a short description since their meaning is not obvious. I think you need a sentence-based description here instead of a bulleted list. How exactly is "personality trait of others" related to RIF categories?
     Resolved

   Can do. The personality trait thing is that the category is a person's name and the exemplars are personality descriptions of that person. This is a little different because participants don't have the benefit of prior knowledge to remember these items as they do with knowing that a banana is a type of fruit, for instance-- they have to learn about this imaginary person all in the context of the study. I, Jethrobot ^{drop me a line} (note: not a bot!)

   • Manipulation of retrieval practice phase: The last two sentences are unsourced. Without having read the sources, I have trouble understanding this section. Fruit-wu should trigger partial retrieval for category fruit? I will look at the sources to see what the picture there is.
     Resolved

   This is an example, though I can pull something more directly form the article cited earlier in the paragraph. I, Jethrobot ^{drop me a line} (note: not a bot!)

   • The two-sentence section should probably be merged with others.  Done
   • Socially shared RIF: I think we don't need the sentence on the benefits of collaborative studying. Instead of an example, might be best to describe "flashbulb memory" directly; the problem with the example is it might convey an idea such memories are related to fear, or worry or whatever; they necessarily aren't.  Done

   Theoretical accounts

   • Interference: For most people, whatever cannot be remembered is not accessible. It might be better to take out "but this does not change their accessibility." That holds only under a restricted definition of the word.  Done
   • Inhibition: "Removing a hand from a hot stove" is a reflex; not an example of inhibition. I don't see the similarity to RIF.  Done

   Ah, good point. I'm falling into a little WP:OR with some of this. I'll just stick with the baseball player example cited in the Anderson (2005) paper. I, Jethrobot ^{drop me a line} (note: not a bot!)

   • Other accounts: Is there a why for this in the sources?
     Resolved

   When you say "why?", to what are you referring, exactly? I, Jethrobot ^{drop me a line} (note: not a bot!)

   I meant why scrambling order changes things; I think it is clearer now. Churn and change (talk) 19:20, 11 November 2012 (UTC)

   B. MoS compliance for lead, layout, words to watch, fiction, and lists:

   • The lead needs to include a statement each on studies in special populations, and socially shared RIF.  Done
   • I think the lead should provide a short explanation of the three-step experimental paradigm rather than using the word "retrieval-practice paradigm" which doesn't convey much to the general reader.  Done
   • The three types of tests also probably merit a one-sentence description in the introduction, since they are prominent in the article, and the lead needs to summarize the article.  Done
   • "When retrieval practice is replaced with additional study": if this section is going to be kept, the title has to change to a noun phrase, in keeping with other titles, say, "Retrieval practice replaced with additional study."  Done
   • I have edited the article to better use the en-dash.  Done

   Thank you, this looks preferable. I, Jethrobot ^{drop me a line} (note: not a bot!)

   • I have also taken out linking words such as "however" at the beginning of a sentence, since that looks like an essay with author linking two separately cited statements.  Done
   • I think the "Main article" links are a bit misleading. The link tells the reader the content ahead is a summary of the main article linked to. That is not true here.  Done

   I can change these to "See also." Would that be more appropriate? The topics are not meant to be full discussions of interference and inhibitory theory, so I thought it would be good to direct them elsewhere for that info. I, Jethrobot ^{drop me a line} (note: not a bot!)

   • Words used as terms (not as standing for their meaning) should be in italics. I notice "Fruit-wu" is italicized, but other cases are not. Note that the MOS can be overridden if you think italics makes reading just harder. But if you want to override it for this situation, you should override it consistently.  Done

   Got it. It was mostly just a stylistic choice to emphasize example stimuli.

   • Now we have citation clutter. It is okay to have two consecutive citation notes, and a rare three. Here now we have four and in one place five. The best option is to just keep the secondary source which says what the statement also says. There are options such as WP:BUNDLING but I think those technical solutions are not the best here. For this statement: "RIF studies have generally yielded results where, on average, unpracticed–related words are remembered less well than the baseline of unpracticed–unrelated words" using five different sources showing the same result to support the "generally" is not okay. If one of those sources does use "generally" or an equivalent, then just citing that source should be sufficient. Churn and change (talk) 21:06, 11 November 2012 (UTC)  Done

   OK, these have been addressed and reduced as needed. Some statements do not have secondary sources, and so I have linked to two empirical studies as supporting verification of those claims. I, Jethrobot ^{drop me a line} (note: not a bot!) 22:02, 11 November 2012 (UTC)

2. Is it factually accurate and verifiable?

   A. References to sources:

   • The references do not need retrieval dates, since, unlike websites, journal articles will not change over time.  Done
   • Links to private copies is problematic since they often end up broken after a while. Best is to link to the doi, stable JSTOR URL, or Google Books.  Done

   OK, I will switch these to dois or JSTORs. I was concerned that without direct links, much of the content would need to be AGF'd on the part of other editors. I, Jethrobot ^{drop me a line} (note: not a bot!)

   • For article title, you need to use either sentence-style (APA) or title-style (MLA) capitalization consistently. Since this is a psychology article, and since you use sentence-style more frequently, might be better to convert all titles to sentence -style caps. Book titles are typically title-style and chapter titles sentence-style. While MOS is not required for GA, consistency would be.  Done
   • A beyond-GA comment would be to move to the {{sfn}} template, which allows multiple citations to be rolled into one note, avoiding clutter. This comment does not have to be taken care of for meeting GA criteria.

   I am not yet familiar with this citation style, but will look into it. I, Jethrobot ^{drop me a line} (note: not a bot!)

   B. Citation of reliable sources where necessary:

   The sources cited—largely credible peer-reviewed journals and books from publishers such as Springer, Academic Press, and the American Psychological Association—are reliable. I have mentioned the few uncited statements in the list above.

   C. No original research:

   Wikipedia requires secondary sources. That would be books, reviews and meta-analyses, and the introductory part of articles where the field is surveyed and summarized. Primary sources, the results of original studies performed in the specific paper, are allowed to some extent. I am a little concerned the balance of primary to secondary sources is a bit out-of-whack here. Storm & Levy (2012), for example is a survey from May this year of the theoretical account (inhibitory account) of RIF, and it is cited just once. One way to handle this is to source as much as possible to the secondary sources; another is to look at publications citing the original source, and see whether the material needed is in the citing publication. I will directly help with this, since GA reviewers are allowed to edit and improve the article. Not asking that all primary sources be removed; just that the balance be shifted more. This source could be used to source the description of the 3-phase experimental paradigm (the description need not change, just the citation).

   Verde (2012) also represents a review of the literature, and there are a few literature reviews I can bring up. There are also a few (although not particularly in-depth) discussions of the phenomenon in popular press that I've been attempting to find and hope to include here. I, Jethrobot ^{drop me a line} (note: not a bot!)

   • The citations to books need page numbers. If the same page number is used for all citations, you can just include the page number in the {{citation}} template. If different page numbers are referred to in different places, you have to use {{sfn}} for those specific cases. For now, you can use the {{rp}} template for such a case, and I will convert it to {{sfn}} Churn and change (talk) 21:10, 11 November 2012 (UTC)  Done

   I only found one book without explicit page numbers in the text (i.e. the MacLeod "In opposition to inhibition" book), and so that has been corrected The other books do not require multiple different sets of page numbers for their citation, at least for now. I, Jethrobot ^{drop me a line} (note: not a bot!) 22:02, 11 November 2012 (UTC)

   • In Bjork described a study by ..., the positions of the citation are reversed. Immediately after "Bjork described a study by X", you should cite Bjork, not X. After you describe what X says, you should cite X, not Bjork.  Done
   • The Brown "Reciprocal Facilitation" paper should not be cited directly because it is a primary source. Considering that it is a 1968 paper cited by over 100 other sources, you could use any of those citing sources as a secondary source. One useful ref usable as a secondary source since it contains no new study or method of calculation would be: Nickerson, R.S. (1984). Retrieval inhibition from part-set cuing: A persisting enigma in memory research. Memory & Cognition, 12(6):531–552. doi:10.3758/BF03213342  Done
   • "Blaxton & Neely, 1983" should be replaced with the secondary source: Anderson, M. C., & Spellman, B. A. (1995). On the status of inhibitory mechanisms in cognition: Memory retrieval as a model case. Psychological Review, 102(1), 68-100. doi:10.1037/0033-295X.102.1.68  Done
   • Response times ranging from 10 seconds to 3 seconds: you need a source which says so. Two sources, one of which has a study with 10 seconds, and another a study with 3 seconds isn't enough, because we don't know whether that covers the whole range.  Done

   I found an article by Camp et al. (2007) that reports this:

   Although a considerable number of studies that demonstrated retrieval-induced forgetting have also used a response time of 10 s (Anderson, Bjork, & Bjork, 1994; Anderson, Green, & McCulloch, 2000; Anderson & Spellman, 1995; Bäuml & Hartinger, 2002; Perfect et al., 2002)...

   The source is here:

   Camp, Gino; Pecher, Diane; Schmidt, & Henk G. No retrieval-induced forgetting using item-specific independent cues: Evidence against a general inhibitory account. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33(5) 950-958. DOI: 10.1037/0278-7393.33.5.950

   I believe this should appropriately support the 10 second commonality. However, I have not found any general discussions of lower response times other than primary sources and sources that vaguely discuss response times in a vague or unspecific manner. I have therefore rephrased this section to still mention that response times vary, but that many studies have retained the 10 second interval. I, Jethrobot ^{drop me a line} (note: not a bot!) 03:13, 14 November 2012 (UTC)

   • RIF studies in special populations: Instead of the primary papers, you could use secondary sources summarizing the work, such as Potss, R., Law, R., Golding, J.F. & Groome, D. (2012). The reliability of retrieval-induced forgetting. European Psychologist 17(1):1–10. doi:10.1027/1016-9040/a000040 which covers individual differences in RIF, depression, dysphoric disorder, schizophrenia, and also reliability and stability of RIF in individuals.  Done
   • Last line of "Inhibition" needs citation.  Done

3. Is it broad in its coverage?

   A. Major aspects:

   The phenomenonis described, how it is studied is detailed, and various related results are covered. A final section explains the theories put forward to explain it. For a GA that is broad-enough coverage. Going beyond GA, I think we should include material on how this is related to the tip-of-the-tongue phenomenon and the like, but this is not required to pass GA.

   B. Focused:

   Article is well-focused on the subject in the title, without meandering into the more general psychological topics of interference, memory, cognition and so forth.

4. Is it neutral?

   Fair representation without bias:

   The phenomenon is considered solid in the literature, and its theoretical underpinnings are still unknown. The article captures these points; the point of view is neutral. On a detail, is the RIF effect study in schizophrenics a widely accepted result? I see a much earlier study show no effect in schizophrenics, but not sure of the nuances there.

   Could you point me to that study? There has not been much work done on schizophrenia in relation to RIF, so depending on the study, it can incorporated as a counterpoint per WP:WEIGHT. I, Jethrobot ^{drop me a line} (note: not a bot!)

   It was a very old study, from 1996 I think, so probably okay ignoring, since newer studies do seem to support no RIF difference for category-cueing but a difference for recognition tests. Churn and change (talk) 21:51, 11 November 2012 (UTC)

5. Is it stable?

   No edit wars, etc:

   Yes. No edit wars, no tags other than easily cleaned dead links; a {{technical}} and a {{RefImprove}} tag in the past, both addressed by now;need to check on one "move" from user-space mentioned in the edit history.

6. Does it contain images to illustrate the topic?

   A. Images are tagged with their copyright status, and valid fair use rationales are provided for non-free content:

   [File:InhibitoryResponse.PNG] seems incorrectly tagged; it is from page 162 of Dynamic Cognitive Processes, editors Ohta and Macleod, 2005, chapter: "The role of inhibitory control in forgetting unwanted memories: A consideration of three methods" by Michael C. Anderson. (see [1]). None of the figures, except for the potted cactus, have a creative element to them and hence aren't copyrightable, but we do need to tag them straight.

   Given that the InhibitoryResponse image is a diagram simple shapes, arrows, and text, is it copyrightable? If it is, would it be possible for me to simply make an alternative image that represents this idea? I, Jethrobot ^{drop me a line} (note: not a bot!)

   Also, I have nominated the problematic image for deletion on Commons. I, Jethrobot ^{drop me a line} (note: not a bot!)

   I would say it is not copyrightable and an alternative image can, and should, be uploaded to WikiCommons. Churn and change (talk) 20:34, 7 November 2012 (UTC)

   B. Images are provided where possible and appropriate, with suitable captions:

   The captions are lucid.

7. Overall:

   Pass or Fail: